;(function (global, factory) {
  typeof exports === 'object' && typeof module !== 'undefined'
    ? factory(exports)
    : typeof define === 'function' && define.amd
    ? define(['exports'], factory)
    : ((global = global || self),
      factory((global.faceapi = global.faceapi || {})))
})(this, function (exports) {
  'use strict'

  /**
   * @license
   * Copyright 2020 Google LLC. All Rights Reserved.
   * Licensed under the Apache License, Version 2.0 (the "License");
   * you may not use this file except in compliance with the License.
   * You may obtain a copy of the License at
   *
   * http://www.apache.org/licenses/LICENSE-2.0
   *
   * Unless required by applicable law or agreed to in writing, software
   * distributed under the License is distributed on an "AS IS" BASIS,
   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   * See the License for the specific language governing permissions and
   * limitations under the License.
   * =============================================================================
   */
  var t = function (e, n) {
    return (t =
      Object.setPrototypeOf ||
      ({ __proto__: [] } instanceof Array &&
        function (t, e) {
          t.__proto__ = e
        }) ||
      function (t, e) {
        for (var n in e) e.hasOwnProperty(n) && (t[n] = e[n])
      })(e, n)
  }
  function e(e, n) {
    function r() {
      this.constructor = e
    }
    t(e, n),
      (e.prototype =
        null === n ? Object.create(n) : ((r.prototype = n.prototype), new r()))
  }
  function n(t, e, n, r) {
    return new (n || (n = Promise))(function (o, a) {
      function i(t) {
        try {
          u(r.next(t))
        } catch (t) {
          a(t)
        }
      }
      function s(t) {
        try {
          u(r.throw(t))
        } catch (t) {
          a(t)
        }
      }
      function u(t) {
        t.done
          ? o(t.value)
          : new n(function (e) {
              e(t.value)
            }).then(i, s)
      }
      u((r = r.apply(t, e || [])).next())
    })
  }
  function r(t, e) {
    var n,
      r,
      o,
      a,
      i = {
        label: 0,
        sent: function () {
          if (1 & o[0]) throw o[1]
          return o[1]
        },
        trys: [],
        ops: [],
      }
    return (
      (a = { next: s(0), throw: s(1), return: s(2) }),
      'function' == typeof Symbol &&
        (a[Symbol.iterator] = function () {
          return this
        }),
      a
    )
    function s(a) {
      return function (s) {
        return (function (a) {
          if (n) throw new TypeError('Generator is already executing.')
          for (; i; )
            try {
              if (
                ((n = 1),
                r &&
                  (o =
                    2 & a[0]
                      ? r.return
                      : a[0]
                      ? r.throw || ((o = r.return) && o.call(r), 0)
                      : r.next) &&
                  !(o = o.call(r, a[1])).done)
              )
                return o
              switch (((r = 0), o && (a = [2 & a[0], o.value]), a[0])) {
                case 0:
                case 1:
                  o = a
                  break
                case 4:
                  return i.label++, { value: a[1], done: !1 }
                case 5:
                  i.label++, (r = a[1]), (a = [0])
                  continue
                case 7:
                  ;(a = i.ops.pop()), i.trys.pop()
                  continue
                default:
                  if (
                    !(o = (o = i.trys).length > 0 && o[o.length - 1]) &&
                    (6 === a[0] || 2 === a[0])
                  ) {
                    i = 0
                    continue
                  }
                  if (3 === a[0] && (!o || (a[1] > o[0] && a[1] < o[3]))) {
                    i.label = a[1]
                    break
                  }
                  if (6 === a[0] && i.label < o[1]) {
                    ;(i.label = o[1]), (o = a)
                    break
                  }
                  if (o && i.label < o[2]) {
                    ;(i.label = o[2]), i.ops.push(a)
                    break
                  }
                  o[2] && i.ops.pop(), i.trys.pop()
                  continue
              }
              a = e.call(t, i)
            } catch (t) {
              ;(a = [6, t]), (r = 0)
            } finally {
              n = o = 0
            }
          if (5 & a[0]) throw a[1]
          return { value: a[0] ? a[1] : void 0, done: !0 }
        })([a, s])
      }
    }
  }
  var o = (function () {
    function t(t) {
      ;(this.global = t),
        (this.flags = {}),
        (this.flagRegistry = {}),
        (this.urlFlags = {}),
        this.populateURLFlags()
    }
    return (
      (t.prototype.setPlatform = function (t, e) {
        null != this.platform &&
          console.warn(
            'Platform ' +
              this.platformName +
              ' has already been set. Overwriting the platform with ' +
              e +
              '.'
          ),
          (this.platformName = t),
          (this.platform = e)
      }),
      (t.prototype.registerFlag = function (t, e, n) {
        if (
          ((this.flagRegistry[t] = { evaluationFn: e, setHook: n }),
          null != this.urlFlags[t])
        ) {
          var r = this.urlFlags[t]
          console.warn(
            'Setting feature override from URL ' + t + ': ' + r + '.'
          ),
            this.set(t, r)
        }
      }),
      (t.prototype.get = function (t) {
        return t in this.flags
          ? this.flags[t]
          : ((this.flags[t] = this.evaluateFlag(t)), this.flags[t])
      }),
      (t.prototype.getNumber = function (t) {
        return this.get(t)
      }),
      (t.prototype.getBool = function (t) {
        return this.get(t)
      }),
      (t.prototype.getFlags = function () {
        return this.flags
      }),
      Object.defineProperty(t.prototype, 'features', {
        get: function () {
          return this.flags
        },
        enumerable: !0,
        configurable: !0,
      }),
      (t.prototype.set = function (t, e) {
        if (null == this.flagRegistry[t])
          throw new Error(
            'Cannot set flag ' + t + ' as it has not been registered.'
          )
        ;(this.flags[t] = e),
          null != this.flagRegistry[t].setHook &&
            this.flagRegistry[t].setHook(e)
      }),
      (t.prototype.evaluateFlag = function (t) {
        if (null == this.flagRegistry[t])
          throw new Error(
            "Cannot evaluate flag '" + t + "': no evaluation function found."
          )
        return this.flagRegistry[t].evaluationFn()
      }),
      (t.prototype.setFlags = function (t) {
        this.flags = Object.assign({}, t)
      }),
      (t.prototype.reset = function () {
        ;(this.flags = {}), (this.urlFlags = {}), this.populateURLFlags()
      }),
      (t.prototype.populateURLFlags = function () {
        var t = this
        if (
          void 0 !== this.global &&
          void 0 !== this.global.location &&
          void 0 !== this.global.location.search
        ) {
          var e,
            n,
            r =
              ((e = this.global.location.search),
              (n = {}),
              e.replace(/[?&]([^=?&]+)(?:=([^&]*))?/g, function (t) {
                for (var e = [], r = 1; r < arguments.length; r++)
                  e[r - 1] = arguments[r]
                return a(n, e[0], e[1]), e.join('=')
              }),
              n)
          if ('tfjsflags' in r)
            r.tfjsflags.split(',').forEach(function (e) {
              var n = e.split(':'),
                r = n[0],
                o = n[1]
              t.urlFlags[r] = (function (t, e) {
                if ('true' === (e = e.toLowerCase()) || 'false' === e)
                  return 'true' === e
                if ('' + +e === e) return +e
                throw new Error(
                  'Could not parse value flag value ' +
                    e +
                    ' for flag ' +
                    t +
                    '.'
                )
              })(r, o)
            })
        }
      }),
      t
    )
  })()
  function a(t, e, n) {
    t[decodeURIComponent(e)] = decodeURIComponent(n || '')
  }
  function i() {
    return s
  }
  var s = null
  var u = new Map(),
    c = new Map()
  function l(t, e) {
    var n = g(t, e)
    return u.get(n)
  }
  function h(t) {
    return c.get(t)
  }
  function f(t) {
    for (var e = u.entries(), n = []; ; ) {
      var r = e.next(),
        o = r.done,
        a = r.value
      if (o) break
      var i = a[0],
        s = a[1]
      i.split('_')[0] === t && n.push(s)
    }
    return n
  }
  function d(t) {
    var e = t.kernelName,
      n = t.backendName,
      r = g(e, n)
    if (u.has(r))
      throw new Error(
        "The kernel '" + e + "' for backend '" + n + "' is already registered"
      )
    u.set(r, t)
  }
  function p(t) {
    var e = t.kernelName
    c.has(e) && console.warn("Overriding the gradient for '" + e + "'"),
      c.set(e, t)
  }
  function v(t, e) {
    var n = g(t, e)
    if (!u.has(n))
      throw new Error(
        "The kernel '" + t + "' for backend '" + e + "' is not registered"
      )
    u.delete(n)
  }
  function m(t) {
    if (!c.has(t))
      throw new Error("The gradient '" + t + "' for backend is not registered")
    c.delete(t)
  }
  function g(t, e) {
    return e + '_' + t
  }
  function y(t) {
    for (var e = t.length, n = 0, r = 0; e > 0; )
      (r = (Math.random() * e) | 0), (n = t[--e]), (t[e] = t[r]), (t[r] = n)
  }
  function x(t, e, n) {
    return Math.max(t, Math.min(e, n))
  }
  function b(t) {
    return t % 2 == 0 ? t : t + 1
  }
  function w(t) {
    for (var e = 0, n = 0; n < t.length; n++) e += t[n]
    return e
  }
  function C(t, e) {
    if (!t) throw new Error('string' == typeof e ? e : e())
  }
  function E(t, e, n) {
    void 0 === n && (n = ''),
      C(S(t, e), function () {
        return n + ' Shapes ' + t + ' and ' + e + ' must match'
      })
  }
  function R(t) {
    C(null != t, function () {
      return 'The input to the tensor constructor must be a non-null value.'
    })
  }
  function I(t, e, n) {
    if (
      (void 0 === e && (e = []),
      void 0 === n && (n = !1),
      null == e && (e = []),
      Array.isArray(t) || (V(t) && !n))
    )
      for (var r = 0; r < t.length; ++r) I(t[r], e, n)
    else e.push(t)
    return e
  }
  function k(t) {
    if (0 === t.length) return 1
    for (var e = t[0], n = 1; n < t.length; n++) e *= t[n]
    return e
  }
  function S(t, e) {
    if (t === e) return !0
    if (null == t || null == e) return !1
    if (t.length !== e.length) return !1
    for (var n = 0; n < t.length; n++) if (t[n] !== e[n]) return !1
    return !0
  }
  function A(t) {
    return t % 1 == 0
  }
  function D(t) {
    if (null != Math.tanh) return Math.tanh(t)
    if (t === 1 / 0) return 1
    if (t === -1 / 0) return -1
    var e = Math.exp(2 * t)
    return (e - 1) / (e + 1)
  }
  function T(t) {
    var e = Math.ceil(Math.sqrt(t))
    return [e, Math.ceil(t / e)]
  }
  function N(t, e) {
    return e <= t.length ? t : t + ' '.repeat(e - t.length)
  }
  function F(t, e, n) {
    return (
      void 0 === e &&
        (e = function (t) {
          return 0
        }),
      new Promise(function (r, o) {
        var a = 0,
          i = function () {
            if (t()) r()
            else {
              a++
              var s = e(a)
              null != n && a >= n ? o() : setTimeout(i, s)
            }
          }
        i()
      })
    )
  }
  function _(t, e) {
    for (var n = 1, r = -1, o = 0; o < t.length; ++o)
      if (t[o] >= 0) n *= t[o]
      else if (-1 === t[o]) {
        if (-1 !== r)
          throw Error(
            'Shapes can only have 1 implicit size. Found -1 at dim ' +
              r +
              ' and dim ' +
              o
          )
        r = o
      } else if (t[o] < 0)
        throw Error('Shapes can not be < 0. Found ' + t[o] + ' at dim ' + o)
    if (-1 === r) {
      if (e > 0 && e !== n)
        throw Error('Size(' + e + ') must match the product of shape ' + t)
      return t
    }
    if (0 === n)
      throw Error(
        'Cannot infer the missing size in [' + t + '] when there are 0 elements'
      )
    if (e % n != 0)
      throw Error(
        "The implicit shape can't be a fractional number. Got " + e + ' / ' + n
      )
    var a = t.slice()
    return (a[r] = e / n), a
  }
  function O(t, e) {
    var n = e.length
    return (
      C(
        (t =
          null == t
            ? e.map(function (t, e) {
                return e
              })
            : [].concat(t)).every(function (t) {
          return t >= -n && t < n
        }),
        function () {
          return (
            'All values in axis param must be in range [-' +
            n +
            ', ' +
            n +
            ') but got axis ' +
            t
          )
        }
      ),
      C(
        t.every(function (t) {
          return A(t)
        }),
        function () {
          return 'All values in axis param must be integers but got axis ' + t
        }
      ),
      t.map(function (t) {
        return t < 0 ? n + t : t
      })
    )
  }
  function M(t, e) {
    for (
      var n = [],
        r = [],
        o = null != e && Array.isArray(e) && 0 === e.length,
        a = null == e || o ? null : O(e, t).sort(),
        i = 0,
        s = 0;
      s < t.length;
      ++s
    ) {
      if (null != a) {
        if (a[i] === s && 1 !== t[s])
          throw new Error(
            "Can't squeeze axis " + s + " since its dim '" + t[s] + "' is not 1"
          )
        ;(null == a[i] || a[i] > s) && 1 === t[s] && (n.push(t[s]), r.push(s)),
          a[i] <= s && i++
      }
      1 !== t[s] && (n.push(t[s]), r.push(s))
    }
    return { newShape: n, keptDims: r }
  }
  function B(t, e) {
    var n = null
    if (null == t || 'float32' === t) n = new Float32Array(e)
    else if ('int32' === t) n = new Int32Array(e)
    else {
      if ('bool' !== t) throw new Error('Unknown data type ' + t)
      n = new Uint8Array(e)
    }
    return n
  }
  function P(t, e) {
    var n = null
    if (null == t || 'float32' === t) n = new Float32Array(e)
    else if ('int32' === t) n = new Int32Array(e)
    else if ('bool' === t) n = new Uint8Array(e)
    else {
      if ('string' !== t) throw new Error('Unknown data type ' + t)
      n = new Array(e)
    }
    return n
  }
  function L(t, e) {
    for (var n = 0; n < t.length; n++) {
      var r = t[n]
      if (isNaN(r) || !isFinite(r))
        throw Error(
          'A tensor of type ' + e + ' being uploaded contains ' + r + '.'
        )
    }
  }
  function W(t) {
    return (
      'bool' === t ||
      'complex64' === t ||
      'float32' === t ||
      'int32' === t ||
      'string' === t
    )
  }
  function U(t, e) {
    return (
      'complex64' !== e &&
      ('float32' !== e || 'complex64' === t) &&
      ('int32' !== e || 'float32' === t || 'complex64' === t) &&
      ('bool' !== e || 'bool' !== t)
    )
  }
  function V(t) {
    return (
      t instanceof Float32Array ||
      t instanceof Int32Array ||
      t instanceof Uint8Array
    )
  }
  function z(t) {
    if ('float32' === t || 'int32' === t) return 4
    if ('complex64' === t) return 8
    if ('bool' === t) return 1
    throw new Error('Unknown dtype ' + t)
  }
  function G(t) {
    if (null == t) return 0
    var e = 0
    return (
      t.forEach(function (t) {
        return (e += t.length)
      }),
      e
    )
  }
  function H(t) {
    return 'string' == typeof t || t instanceof String
  }
  function q(t) {
    return 'boolean' == typeof t
  }
  function K(t) {
    return 'number' == typeof t
  }
  function j(t) {
    return Array.isArray(t)
      ? j(t[0])
      : t instanceof Float32Array
      ? 'float32'
      : t instanceof Int32Array || t instanceof Uint8Array
      ? 'int32'
      : K(t)
      ? 'float32'
      : H(t)
      ? 'string'
      : q(t)
      ? 'bool'
      : 'float32'
  }
  function X(t) {
    return !!(t && t.constructor && t.call && t.apply)
  }
  function Y(t, e) {
    for (var n = e; n < t; ++n) if (t % n == 0) return n
    return t
  }
  function $(t) {
    var e = t.length
    if (e < 2) return []
    var n = new Array(e - 1)
    n[e - 2] = t[e - 1]
    for (var r = e - 3; r >= 0; --r) n[r] = n[r + 1] * t[r + 1]
    return n
  }
  function Q(t, e, n) {
    if ('string' === e)
      throw new Error('Cannot convert a string[] to a TypedArray')
    if (
      (Array.isArray(t) && (t = I(t)),
      n && L(t, e),
      (function (t, e) {
        return (
          (t instanceof Float32Array && 'float32' === e) ||
          (t instanceof Int32Array && 'int32' === e) ||
          (t instanceof Uint8Array && 'bool' === e)
        )
      })(t, e))
    )
      return t
    if (null == e || 'float32' === e || 'complex64' === e)
      return new Float32Array(t)
    if ('int32' === e) return new Int32Array(t)
    if ('bool' === e) {
      for (var r = new Uint8Array(t.length), o = 0; o < r.length; ++o)
        0 !== Math.round(t[o]) && (r[o] = 1)
      return r
    }
    throw new Error('Unknown data type ' + e)
  }
  function J(t, e) {
    if (0 === t.length) return e[0]
    var n = t.reduce(function (t, e) {
      return t * e
    })
    if (0 === n) return []
    if (n !== e.length)
      throw new Error('[' + t + '] does not match the input size.')
    return (function t(e, n, r) {
      var o = new Array()
      if (1 === n.length) for (var a = n[0], i = 0; i < a; i++) o[i] = r[e + i]
      else {
        a = n[0]
        var s = n.slice(1),
          u = s.reduce(function (t, e) {
            return t * e
          })
        for (i = 0; i < a; i++) o[i] = t(e + i * u, s, r)
      }
      return o
    })(0, t, e)
  }
  function Z(t, e) {
    for (var n = tt(t, e), r = 0; r < n.length; r++) n[r] = 1
    return n
  }
  function tt(t, e) {
    if (null == e || 'float32' === e || 'complex64' === e)
      return new Float32Array(t)
    if ('int32' === e) return new Int32Array(t)
    if ('bool' === e) return new Uint8Array(t)
    throw new Error('Unknown data type ' + e)
  }
  function et() {
    return i().platform.now()
  }
  function nt(t) {
    t.forEach(function (e) {
      C(Number.isInteger(e) && e >= 0, function () {
        return (
          'Tensor must have a shape comprised of positive integers but got shape [' +
          t +
          '].'
        )
      })
    })
  }
  function rt(t, e) {
    return (
      void 0 === e && (e = 'utf-8'),
      (e = e || 'utf-8'),
      i().platform.encode(t, e)
    )
  }
  function ot(t, e) {
    return (
      void 0 === e && (e = 'utf-8'),
      (e = e || 'utf-8'),
      i().platform.decode(t, e)
    )
  }
  function at(t, e, n) {
    if (0 === e) return 0
    if (1 === e) return t[0]
    for (var r = t[t.length - 1], o = 0; o < t.length - 1; ++o) r += n[o] * t[o]
    return r
  }
  function it(t, e, n) {
    if (0 === e) return []
    if (1 === e) return [t]
    for (var r = new Array(e), o = 0; o < r.length - 1; ++o)
      (r[o] = Math.floor(t / n[o])), (t -= r[o] * n[o])
    return (r[r.length - 1] = t), r
  }
  var st = Object.freeze({
      shuffle: y,
      clamp: x,
      nearestLargerEven: b,
      sum: w,
      randUniform: function (t, e) {
        var n = Math.random()
        return e * n + (1 - n) * t
      },
      distSquared: function (t, e) {
        for (var n = 0, r = 0; r < t.length; r++) {
          var o = Number(t[r]) - Number(e[r])
          n += o * o
        }
        return n
      },
      assert: C,
      assertShapesMatch: E,
      assertNonNull: R,
      flatten: I,
      sizeFromShape: k,
      isScalarShape: function (t) {
        return 0 === t.length
      },
      arraysEqual: S,
      isInt: A,
      tanh: D,
      sizeToSquarishShape: T,
      createShuffledIndices: function (t) {
        for (var e = new Uint32Array(t), n = 0; n < t; ++n) e[n] = n
        return y(e), e
      },
      rightPad: N,
      repeatedTry: F,
      inferFromImplicitShape: _,
      parseAxisParam: O,
      squeezeShape: M,
      getTypedArrayFromDType: B,
      getArrayFromDType: P,
      checkConversionForErrors: L,
      isValidDtype: W,
      hasEncodingLoss: U,
      isTypedArray: V,
      bytesPerElement: z,
      bytesFromStringArray: G,
      isString: H,
      isBoolean: q,
      isNumber: K,
      inferDtype: j,
      isFunction: X,
      nearestDivisor: Y,
      computeStrides: $,
      toTypedArray: Q,
      toNestedArray: J,
      makeOnesTypedArray: Z,
      makeZerosTypedArray: tt,
      now: et,
      assertNonNegativeIntegerDimensions: nt,
      fetch: function (t, e) {
        return i().platform.fetch(t, e)
      },
      encodeString: rt,
      decodeString: ot,
      locToIndex: at,
      indexToLoc: it,
    }),
    ut = (function () {
      function t(t, e) {
        ;(this.backendTimer = t),
          (this.logger = e),
          null == e && (this.logger = new ct())
      }
      return (
        (t.prototype.profileKernel = function (t, e, n) {
          var r,
            o = this,
            a = this.backendTimer.time(function () {
              r = n()
            })
          return (
            r.forEach(function (n) {
              n.data().then(function (r) {
                !(function (t, e, n) {
                  if ('float32' !== e) return !1
                  for (var r = 0; r < t.length; r++) {
                    var o = t[r]
                    if (isNaN(o) || !isFinite(o))
                      return (
                        console.warn(
                          'Found ' + o + " in the result of '" + n + "'"
                        ),
                        !0
                      )
                  }
                })(r, n.dtype, t),
                  a.then(function (a) {
                    var i = ''
                    null != a.getExtraProfileInfo &&
                      (i = a.getExtraProfileInfo()),
                      o.logger.logKernelProfile(t, n, r, a.kernelMs, e, i)
                  })
              })
            }),
            r
          )
        }),
        t
      )
    })()
  var ct = (function () {
    function t() {}
    return (
      (t.prototype.logKernelProfile = function (t, e, n, r, o, a) {
        var i = 'number' == typeof r ? N(r + 'ms', 9) : r.error,
          s = N(t, 25),
          u = e.rank,
          c = e.size,
          l = N(e.shape.toString(), 14),
          h = ''
        for (var f in o) {
          var d = o[f].shape || e.shape,
            p = d.length
          h += f + ': ' + p + 'D ' + (p > 0 ? d : '') + ' '
        }
        console.log(
          '%c' +
            s +
            '\t%c' +
            i +
            '\t%c' +
            u +
            'D ' +
            l +
            '\t%c' +
            c +
            '\t%c' +
            h +
            '\t%c' +
            a,
          'font-weight:bold',
          'color:red',
          'color:blue',
          'color: orange',
          'color: green',
          'color: steelblue'
        )
      }),
      t
    )
  })()
  var lt = 20,
    ht = 3,
    ft = 7
  function dt(t, e, n, r) {
    var o = $(e),
      a = (function (t, e, n, r) {
        var o = k(e),
          a = r[r.length - 1],
          i = new Array(a).fill(0),
          s = e.length,
          u = 'complex64' === n ? mt(t) : t
        if (s > 1)
          for (var c = 0; c < o / a; c++)
            for (var l = c * a, h = 0; h < a; h++)
              i[h] = Math.max(i[h], pt(u[l + h], 0, n).length)
        return i
      })(t, e, n, o),
      i = e.length,
      s = (function t(e, n, r, o, a, i) {
        void 0 === i && (i = !0)
        var s = 'complex64' === r ? 2 : 1,
          u = n[0],
          c = n.length
        if (0 === c) {
          return 'complex64' === r
            ? [pt(mt(e)[0], 0, r)]
            : 'bool' === r
            ? [vt(e[0])]
            : [e[0].toString()]
        }
        if (1 === c) {
          if (u > lt) {
            var l = ht * s,
              h = Array.from(e.slice(0, l)),
              f = Array.from(e.slice((u - ht) * s, u * s))
            return (
              'complex64' === r && ((h = mt(h)), (f = mt(f))),
              [
                '[' +
                  h
                    .map(function (t, e) {
                      return pt(t, a[e], r)
                    })
                    .join(', ') +
                  ', ..., ' +
                  f
                    .map(function (t, e) {
                      return pt(t, a[u - ht + e], r)
                    })
                    .join(', ') +
                  ']',
              ]
            )
          }
          return [
            '[' +
              ('complex64' === r ? mt(e) : Array.from(e))
                .map(function (t, e) {
                  return pt(t, a[e], r)
                })
                .join(', ') +
              ']',
          ]
        }
        var d = n.slice(1),
          p = o.slice(1),
          v = o[0] * s,
          m = []
        if (u > lt) {
          for (var g = 0; g < ht; g++) {
            var y = (x = g * v) + v
            m.push.apply(m, t(e.slice(x, y), d, r, p, a, !1))
          }
          m.push('...')
          for (g = u - ht; g < u; g++) {
            y = (x = g * v) + v
            m.push.apply(m, t(e.slice(x, y), d, r, p, a, g === u - 1))
          }
        } else
          for (g = 0; g < u; g++) {
            var x
            y = (x = g * v) + v
            m.push.apply(m, t(e.slice(x, y), d, r, p, a, g === u - 1))
          }
        var b = 2 === c ? ',' : ''
        m[0] = '[' + m[0] + b
        for (g = 1; g < m.length - 1; g++) m[g] = ' ' + m[g] + b
        var w = ',\n'
        for (g = 2; g < c; g++) w += '\n'
        return (m[m.length - 1] = ' ' + m[m.length - 1] + ']' + (i ? '' : w)), m
      })(t, e, n, o, a),
      u = ['Tensor']
    return (
      r &&
        (u.push('  dtype: ' + n),
        u.push('  rank: ' + i),
        u.push('  shape: [' + e + ']'),
        u.push('  values:')),
      u.push(
        s
          .map(function (t) {
            return '    ' + t
          })
          .join('\n')
      ),
      u.join('\n')
    )
  }
  function pt(t, e, n) {
    return N(
      Array.isArray(t)
        ? parseFloat(t[0].toFixed(ft)) +
            ' + ' +
            parseFloat(t[1].toFixed(ft)) +
            'j'
        : H(t)
        ? "'" + t + "'"
        : 'bool' === n
        ? vt(t)
        : parseFloat(t.toFixed(ft)).toString(),
      e
    )
  }
  function vt(t) {
    return 0 === t ? 'false' : 'true'
  }
  function mt(t) {
    for (var e = [], n = 0; n < t.length; n += 2) e.push([t[n], t[n + 1]])
    return e
  }
  var gt = (function () {
      function t(t, e, n) {
        var r = this
        if (
          ((this.dtype = e),
          (this.shape = t.slice()),
          (this.size = k(t)),
          null != n)
        ) {
          var o = n.length
          C(o === this.size, function () {
            return (
              "Length of values '" +
              o +
              "' does not match the size inferred by the shape '" +
              r.size +
              "'."
            )
          })
        }
        if ('complex64' === e)
          throw new Error(
            'complex64 dtype TensorBuffers are not supported. Please create a TensorBuffer for the real and imaginary parts separately and call tf.complex(real, imag).'
          )
        ;(this.values = n || P(e, this.size)), (this.strides = $(t))
      }
      return (
        (t.prototype.set = function (t) {
          for (var e = this, n = [], r = 1; r < arguments.length; r++)
            n[r - 1] = arguments[r]
          0 === n.length && (n = [0]),
            C(n.length === this.rank, function () {
              return (
                'The number of provided coordinates (' +
                n.length +
                ') must match the rank (' +
                e.rank +
                ')'
              )
            })
          var o = this.locToIndex(n)
          this.values[o] = t
        }),
        (t.prototype.get = function () {
          for (var t = [], e = 0; e < arguments.length; e++) t[e] = arguments[e]
          0 === t.length && (t = [0])
          for (var n = 0, r = 0, o = t; r < o.length; r++) {
            var a = o[r]
            if (a < 0 || a >= this.shape[n]) {
              var i =
                'Requested out of range element at ' +
                t +
                '.   Buffer shape=' +
                this.shape
              throw new Error(i)
            }
            n++
          }
          for (var s = t[t.length - 1], u = 0; u < t.length - 1; ++u)
            s += this.strides[u] * t[u]
          return this.values[s]
        }),
        (t.prototype.locToIndex = function (t) {
          if (0 === this.rank) return 0
          if (1 === this.rank) return t[0]
          for (var e = t[t.length - 1], n = 0; n < t.length - 1; ++n)
            e += this.strides[n] * t[n]
          return e
        }),
        (t.prototype.indexToLoc = function (t) {
          if (0 === this.rank) return []
          if (1 === this.rank) return [t]
          for (
            var e = new Array(this.shape.length), n = 0;
            n < e.length - 1;
            ++n
          )
            (e[n] = Math.floor(t / this.strides[n])),
              (t -= e[n] * this.strides[n])
          return (e[e.length - 1] = t), e
        }),
        Object.defineProperty(t.prototype, 'rank', {
          get: function () {
            return this.shape.length
          },
          enumerable: !0,
          configurable: !0,
        }),
        (t.prototype.toTensor = function () {
          return yt().makeTensor(this.values, this.shape, this.dtype)
        }),
        t
      )
    })(),
    yt = null,
    xt = null,
    bt = null
  var wt = (function () {
    function t(t, e, n, r) {
      ;(this.kept = !1),
        (this.isDisposedInternal = !1),
        (this.shape = t.slice()),
        (this.dtype = e || 'float32'),
        (this.size = k(t)),
        (this.strides = $(t)),
        (this.dataId = n),
        (this.id = r),
        (this.rankType = this.rank < 5 ? this.rank.toString() : 'higher')
    }
    return (
      (t.prototype.flatten = function () {
        return this.throwIfDisposed(), this.as1D()
      }),
      (t.prototype.asScalar = function () {
        return (
          this.throwIfDisposed(),
          C(1 === this.size, function () {
            return 'The array must have only 1 element.'
          }),
          this.reshape([])
        )
      }),
      (t.prototype.as1D = function () {
        return this.throwIfDisposed(), this.reshape([this.size])
      }),
      (t.prototype.as2D = function (t, e) {
        return this.throwIfDisposed(), this.reshape([t, e])
      }),
      (t.prototype.as3D = function (t, e, n) {
        return this.throwIfDisposed(), this.reshape([t, e, n])
      }),
      (t.prototype.as4D = function (t, e, n, r) {
        return this.throwIfDisposed(), this.reshape([t, e, n, r])
      }),
      (t.prototype.as5D = function (t, e, n, r, o) {
        return this.throwIfDisposed(), this.reshape([t, e, n, r, o])
      }),
      (t.prototype.asType = function (t) {
        return this.throwIfDisposed(), xt.cast(this, t)
      }),
      Object.defineProperty(t.prototype, 'rank', {
        get: function () {
          return this.shape.length
        },
        enumerable: !0,
        configurable: !0,
      }),
      (t.prototype.buffer = function () {
        return n(this, void 0, void 0, function () {
          var t
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return [4, this.data()]
              case 1:
                return (t = e.sent()), [2, xt.buffer(this.shape, this.dtype, t)]
            }
          })
        })
      }),
      (t.prototype.bufferSync = function () {
        return xt.buffer(this.shape, this.dtype, this.dataSync())
      }),
      (t.prototype.array = function () {
        return n(this, void 0, void 0, function () {
          var t
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return [4, this.data()]
              case 1:
                return (t = e.sent()), [2, J(this.shape, t)]
            }
          })
        })
      }),
      (t.prototype.arraySync = function () {
        return J(this.shape, this.dataSync())
      }),
      (t.prototype.data = function () {
        return n(this, void 0, void 0, function () {
          var t, e
          return r(this, function (n) {
            switch (n.label) {
              case 0:
                return (
                  this.throwIfDisposed(),
                  (t = yt().read(this.dataId)),
                  'string' !== this.dtype ? [3, 2] : [4, t]
                )
              case 1:
                e = n.sent()
                try {
                  return [
                    2,
                    e.map(function (t) {
                      return ot(t)
                    }),
                  ]
                } catch (t) {
                  throw new Error(
                    'Failed to decode the string bytes into utf-8. To get the original bytes, call tensor.bytes().'
                  )
                }
                n.label = 2
              case 2:
                return [2, t]
            }
          })
        })
      }),
      (t.prototype.dataSync = function () {
        this.throwIfDisposed()
        var t = yt().readSync(this.dataId)
        if ('string' === this.dtype)
          try {
            return t.map(function (t) {
              return ot(t)
            })
          } catch (t) {
            throw new Error(
              'Failed to decode the string bytes into utf-8. To get the original bytes, call tensor.bytes().'
            )
          }
        return t
      }),
      (t.prototype.bytes = function () {
        return n(this, void 0, void 0, function () {
          var t
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return this.throwIfDisposed(), [4, yt().read(this.dataId)]
              case 1:
                return (
                  (t = e.sent()),
                  'string' === this.dtype
                    ? [2, t]
                    : [2, new Uint8Array(t.buffer)]
                )
            }
          })
        })
      }),
      (t.prototype.dispose = function () {
        this.isDisposed ||
          (yt().disposeTensor(this), (this.isDisposedInternal = !0))
      }),
      Object.defineProperty(t.prototype, 'isDisposed', {
        get: function () {
          return this.isDisposedInternal
        },
        enumerable: !0,
        configurable: !0,
      }),
      (t.prototype.throwIfDisposed = function () {
        if (this.isDisposed) throw new Error('Tensor is disposed.')
      }),
      (t.prototype.toFloat = function () {
        return this.asType('float32')
      }),
      (t.prototype.toInt = function () {
        return this.asType('int32')
      }),
      (t.prototype.toBool = function () {
        return this.asType('bool')
      }),
      (t.prototype.print = function (t) {
        return void 0 === t && (t = !1), xt.print(this, t)
      }),
      (t.prototype.reshape = function (t) {
        return this.throwIfDisposed(), xt.reshape(this, t)
      }),
      (t.prototype.reshapeAs = function (t) {
        return this.throwIfDisposed(), this.reshape(t.shape)
      }),
      (t.prototype.expandDims = function (t) {
        return void 0 === t && (t = 0), xt.expandDims(this, t)
      }),
      (t.prototype.cumsum = function (t, e, n) {
        return (
          void 0 === t && (t = 0),
          void 0 === e && (e = !1),
          void 0 === n && (n = !1),
          xt.cumsum(this, t, e, n)
        )
      }),
      (t.prototype.squeeze = function (t) {
        return this.throwIfDisposed(), xt.squeeze(this, t)
      }),
      (t.prototype.clone = function () {
        return this.throwIfDisposed(), xt.clone(this)
      }),
      (t.prototype.oneHot = function (t, e, n) {
        return this.throwIfDisposed(), xt.oneHot(this, t, e, n)
      }),
      (t.prototype.toString = function (t) {
        return (
          void 0 === t && (t = !1),
          dt(this.dataSync(), this.shape, this.dtype, t)
        )
      }),
      (t.prototype.tile = function (t) {
        return this.throwIfDisposed(), xt.tile(this, t)
      }),
      (t.prototype.gather = function (t, e) {
        return (
          void 0 === e && (e = 0), this.throwIfDisposed(), xt.gather(this, t, e)
        )
      }),
      (t.prototype.matMul = function (t, e, n) {
        return (
          void 0 === e && (e = !1),
          void 0 === n && (n = !1),
          this.throwIfDisposed(),
          xt.matMul(this, t, e, n)
        )
      }),
      (t.prototype.dot = function (t) {
        return this.throwIfDisposed(), xt.dot(this, t)
      }),
      (t.prototype.norm = function (t, e, n) {
        return (
          void 0 === t && (t = 'euclidean'),
          void 0 === e && (e = null),
          void 0 === n && (n = !1),
          this.throwIfDisposed(),
          xt.norm(this, t, e, n)
        )
      }),
      (t.prototype.slice = function (t, e) {
        return this.throwIfDisposed(), xt.slice(this, t, e)
      }),
      (t.prototype.reverse = function (t) {
        return this.throwIfDisposed(), xt.reverse(this, t)
      }),
      (t.prototype.concat = function (e, n) {
        return (
          void 0 === n && (n = 0),
          this.throwIfDisposed(),
          e instanceof t && (e = [e]),
          xt.concat([this].concat(e), n)
        )
      }),
      (t.prototype.split = function (t, e) {
        return (
          void 0 === e && (e = 0), this.throwIfDisposed(), xt.split(this, t, e)
        )
      }),
      (t.prototype.stack = function (t, e) {
        return void 0 === e && (e = 0), xt.stack([this, t], e)
      }),
      (t.prototype.unstack = function (t) {
        return void 0 === t && (t = 0), xt.unstack(this, t)
      }),
      (t.prototype.pad = function (t, e) {
        return void 0 === e && (e = 0), xt.pad(this, t, e)
      }),
      (t.prototype.batchNormalization = function (t, e, n, r, o) {
        return (
          void 0 === n && (n = 0.001),
          bt(
            'tf.batchNormalization() is going away. Use tf.batchNorm() instead, and note the positional argument change of scale, offset, and varianceEpsilon'
          ),
          this.batchNorm(t, e, o, r, n)
        )
      }),
      (t.prototype.batchNorm = function (t, e, n, r, o) {
        return (
          void 0 === o && (o = 0.001),
          this.throwIfDisposed(),
          xt.batchNorm(this, t, e, n, r, o)
        )
      }),
      (t.prototype.all = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.all(this, t, e)
        )
      }),
      (t.prototype.any = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.any(this, t, e)
        )
      }),
      (t.prototype.logSumExp = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.logSumExp(this, t, e)
        )
      }),
      (t.prototype.sum = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.sum(this, t, e)
        )
      }),
      (t.prototype.prod = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.prod(this, t, e)
        )
      }),
      (t.prototype.mean = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.mean(this, t, e)
        )
      }),
      (t.prototype.min = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.min(this, t, e)
        )
      }),
      (t.prototype.max = function (t, e) {
        return (
          void 0 === t && (t = null),
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.max(this, t, e)
        )
      }),
      (t.prototype.argMin = function (t) {
        return (
          void 0 === t && (t = null), this.throwIfDisposed(), xt.argMin(this, t)
        )
      }),
      (t.prototype.argMax = function (t) {
        return (
          void 0 === t && (t = null), this.throwIfDisposed(), xt.argMax(this, t)
        )
      }),
      (t.prototype.cast = function (t) {
        return this.throwIfDisposed(), xt.cast(this, t)
      }),
      (t.prototype.add = function (t) {
        return this.throwIfDisposed(), xt.add(this, t)
      }),
      (t.prototype.addStrict = function (t) {
        return this.throwIfDisposed(), xt.addStrict(this, t)
      }),
      (t.prototype.atan2 = function (t) {
        return this.throwIfDisposed(), xt.atan2(this, t)
      }),
      (t.prototype.sub = function (t) {
        return this.throwIfDisposed(), xt.sub(this, t)
      }),
      (t.prototype.subStrict = function (t) {
        return this.throwIfDisposed(), xt.subStrict(this, t)
      }),
      (t.prototype.pow = function (t) {
        return this.throwIfDisposed(), xt.pow(this, t)
      }),
      (t.prototype.powStrict = function (t) {
        return this.throwIfDisposed(), xt.powStrict(this, t)
      }),
      (t.prototype.mul = function (t) {
        return this.throwIfDisposed(), xt.mul(this, t)
      }),
      (t.prototype.mulStrict = function (t) {
        return this.throwIfDisposed(), xt.mulStrict(this, t)
      }),
      (t.prototype.div = function (t) {
        return this.throwIfDisposed(), xt.div(this, t)
      }),
      (t.prototype.divNoNan = function (t) {
        return this.throwIfDisposed(), xt.divNoNan(this, t)
      }),
      (t.prototype.floorDiv = function (t) {
        return this.throwIfDisposed(), xt.floorDiv(this, t)
      }),
      (t.prototype.divStrict = function (t) {
        return this.throwIfDisposed(), xt.divStrict(this, t)
      }),
      (t.prototype.minimum = function (t) {
        return this.throwIfDisposed(), xt.minimum(this, t)
      }),
      (t.prototype.minimumStrict = function (t) {
        return this.throwIfDisposed(), xt.minimumStrict(this, t)
      }),
      (t.prototype.maximum = function (t) {
        return this.throwIfDisposed(), xt.maximum(this, t)
      }),
      (t.prototype.maximumStrict = function (t) {
        return this.throwIfDisposed(), xt.maximumStrict(this, t)
      }),
      (t.prototype.mod = function (t) {
        return this.throwIfDisposed(), xt.mod(this, t)
      }),
      (t.prototype.modStrict = function (t) {
        return this.throwIfDisposed(), xt.modStrict(this, t)
      }),
      (t.prototype.squaredDifferenceStrict = function (t) {
        return this.throwIfDisposed(), xt.squaredDifferenceStrict(this, t)
      }),
      (t.prototype.transpose = function (t) {
        return this.throwIfDisposed(), xt.transpose(this, t)
      }),
      (t.prototype.notEqual = function (t) {
        return this.throwIfDisposed(), xt.notEqual(this, t)
      }),
      (t.prototype.notEqualStrict = function (t) {
        return this.throwIfDisposed(), xt.notEqualStrict(this, t)
      }),
      (t.prototype.less = function (t) {
        return this.throwIfDisposed(), xt.less(this, t)
      }),
      (t.prototype.lessStrict = function (t) {
        return this.throwIfDisposed(), xt.lessStrict(this, t)
      }),
      (t.prototype.equal = function (t) {
        return this.throwIfDisposed(), xt.equal(this, t)
      }),
      (t.prototype.equalStrict = function (t) {
        return this.throwIfDisposed(), xt.equalStrict(this, t)
      }),
      (t.prototype.lessEqual = function (t) {
        return this.throwIfDisposed(), xt.lessEqual(this, t)
      }),
      (t.prototype.lessEqualStrict = function (t) {
        return this.throwIfDisposed(), xt.lessEqualStrict(this, t)
      }),
      (t.prototype.greater = function (t) {
        return this.throwIfDisposed(), xt.greater(this, t)
      }),
      (t.prototype.greaterStrict = function (t) {
        return this.throwIfDisposed(), xt.greaterStrict(this, t)
      }),
      (t.prototype.greaterEqual = function (t) {
        return this.throwIfDisposed(), xt.greaterEqual(this, t)
      }),
      (t.prototype.greaterEqualStrict = function (t) {
        return this.throwIfDisposed(), xt.greaterEqualStrict(this, t)
      }),
      (t.prototype.logicalAnd = function (t) {
        return this.throwIfDisposed(), xt.logicalAnd(this, t)
      }),
      (t.prototype.logicalOr = function (t) {
        return this.throwIfDisposed(), xt.logicalOr(this, t)
      }),
      (t.prototype.logicalNot = function () {
        return this.throwIfDisposed(), xt.logicalNot(this)
      }),
      (t.prototype.logicalXor = function (t) {
        return this.throwIfDisposed(), xt.logicalXor(this, t)
      }),
      (t.prototype.where = function (t, e) {
        return this.throwIfDisposed(), xt.where(t, this, e)
      }),
      (t.prototype.neg = function () {
        return this.throwIfDisposed(), xt.neg(this)
      }),
      (t.prototype.ceil = function () {
        return this.throwIfDisposed(), xt.ceil(this)
      }),
      (t.prototype.floor = function () {
        return this.throwIfDisposed(), xt.floor(this)
      }),
      (t.prototype.sign = function () {
        return this.throwIfDisposed(), xt.sign(this)
      }),
      (t.prototype.isNaN = function () {
        return this.throwIfDisposed(), xt.isNaN(this)
      }),
      (t.prototype.isInf = function () {
        return this.throwIfDisposed(), xt.isInf(this)
      }),
      (t.prototype.isFinite = function () {
        return this.throwIfDisposed(), xt.isFinite(this)
      }),
      (t.prototype.exp = function () {
        return this.throwIfDisposed(), xt.exp(this)
      }),
      (t.prototype.expm1 = function () {
        return this.throwIfDisposed(), xt.expm1(this)
      }),
      (t.prototype.log = function () {
        return this.throwIfDisposed(), xt.log(this)
      }),
      (t.prototype.log1p = function () {
        return this.throwIfDisposed(), xt.log1p(this)
      }),
      (t.prototype.sqrt = function () {
        return this.throwIfDisposed(), xt.sqrt(this)
      }),
      (t.prototype.rsqrt = function () {
        return this.throwIfDisposed(), xt.rsqrt(this)
      }),
      (t.prototype.square = function () {
        return this.throwIfDisposed(), xt.square(this)
      }),
      (t.prototype.reciprocal = function () {
        return this.throwIfDisposed(), xt.reciprocal(this)
      }),
      (t.prototype.abs = function () {
        return this.throwIfDisposed(), xt.abs(this)
      }),
      (t.prototype.clipByValue = function (t, e) {
        return this.throwIfDisposed(), xt.clipByValue(this, t, e)
      }),
      (t.prototype.relu = function () {
        return this.throwIfDisposed(), xt.relu(this)
      }),
      (t.prototype.relu6 = function () {
        return this.throwIfDisposed(), xt.relu6(this)
      }),
      (t.prototype.elu = function () {
        return this.throwIfDisposed(), xt.elu(this)
      }),
      (t.prototype.selu = function () {
        return this.throwIfDisposed(), xt.selu(this)
      }),
      (t.prototype.leakyRelu = function (t) {
        return (
          void 0 === t && (t = 0.2),
          this.throwIfDisposed(),
          xt.leakyRelu(this, t)
        )
      }),
      (t.prototype.prelu = function (t) {
        return this.throwIfDisposed(), xt.prelu(this, t)
      }),
      (t.prototype.sigmoid = function () {
        return this.throwIfDisposed(), xt.sigmoid(this)
      }),
      (t.prototype.logSigmoid = function () {
        return this.throwIfDisposed(), xt.logSigmoid(this)
      }),
      (t.prototype.softplus = function () {
        return this.throwIfDisposed(), xt.softplus(this)
      }),
      (t.prototype.zerosLike = function () {
        return this.throwIfDisposed(), xt.zerosLike(this)
      }),
      (t.prototype.onesLike = function () {
        return this.throwIfDisposed(), xt.onesLike(this)
      }),
      (t.prototype.sin = function () {
        return this.throwIfDisposed(), xt.sin(this)
      }),
      (t.prototype.cos = function () {
        return this.throwIfDisposed(), xt.cos(this)
      }),
      (t.prototype.tan = function () {
        return this.throwIfDisposed(), xt.tan(this)
      }),
      (t.prototype.asin = function () {
        return this.throwIfDisposed(), xt.asin(this)
      }),
      (t.prototype.acos = function () {
        return this.throwIfDisposed(), xt.acos(this)
      }),
      (t.prototype.atan = function () {
        return this.throwIfDisposed(), xt.atan(this)
      }),
      (t.prototype.sinh = function () {
        return this.throwIfDisposed(), xt.sinh(this)
      }),
      (t.prototype.cosh = function () {
        return this.throwIfDisposed(), xt.cosh(this)
      }),
      (t.prototype.tanh = function () {
        return this.throwIfDisposed(), xt.tanh(this)
      }),
      (t.prototype.asinh = function () {
        return this.throwIfDisposed(), xt.asinh(this)
      }),
      (t.prototype.acosh = function () {
        return this.throwIfDisposed(), xt.acosh(this)
      }),
      (t.prototype.atanh = function () {
        return this.throwIfDisposed(), xt.atanh(this)
      }),
      (t.prototype.erf = function () {
        return this.throwIfDisposed(), xt.erf(this)
      }),
      (t.prototype.round = function () {
        return this.throwIfDisposed(), xt.round(this)
      }),
      (t.prototype.step = function (t) {
        return void 0 === t && (t = 0), this.throwIfDisposed(), xt.step(this, t)
      }),
      (t.prototype.softmax = function (t) {
        return (
          void 0 === t && (t = -1), this.throwIfDisposed(), xt.softmax(this, t)
        )
      }),
      (t.prototype.logSoftmax = function (t) {
        return (
          void 0 === t && (t = -1),
          this.throwIfDisposed(),
          xt.logSoftmax(this, t)
        )
      }),
      (t.prototype.resizeBilinear = function (t, e) {
        return (
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.image.resizeBilinear(this, t, e)
        )
      }),
      (t.prototype.resizeNearestNeighbor = function (t, e) {
        return (
          void 0 === e && (e = !1),
          this.throwIfDisposed(),
          xt.image.resizeNearestNeighbor(this, t, e)
        )
      }),
      (t.prototype.conv1d = function (t, e, n, r, o, a) {
        return (
          void 0 === r && (r = 'NWC'),
          void 0 === o && (o = 1),
          this.throwIfDisposed(),
          xt.conv1d(this, t, e, n, r, o, a)
        )
      }),
      (t.prototype.conv2d = function (t, e, n, r, o, a) {
        return (
          void 0 === r && (r = 'NHWC'),
          void 0 === o && (o = [1, 1]),
          this.throwIfDisposed(),
          xt.conv2d(this, t, e, n, r, o, a)
        )
      }),
      (t.prototype.conv2dTranspose = function (t, e, n, r, o) {
        return this.throwIfDisposed(), xt.conv2dTranspose(this, t, e, n, r, o)
      }),
      (t.prototype.depthwiseConv2D = function (t, e, n, r, o, a) {
        return (
          void 0 === r && (r = 'NHWC'),
          void 0 === o && (o = [1, 1]),
          this.throwIfDisposed(),
          xt.depthwiseConv2d(this, t, e, n, r, o, a)
        )
      }),
      (t.prototype.separableConv2d = function (t, e, n, r, o, a) {
        return (
          void 0 === o && (o = [1, 1]),
          void 0 === a && (a = 'NHWC'),
          this.throwIfDisposed(),
          xt.separableConv2d(this, t, e, n, r, o, a)
        )
      }),
      (t.prototype.avgPool = function (t, e, n, r) {
        return this.throwIfDisposed(), xt.avgPool(this, t, e, n, r)
      }),
      (t.prototype.maxPool = function (t, e, n, r) {
        return this.throwIfDisposed(), xt.maxPool(this, t, e, n, r)
      }),
      (t.prototype.localResponseNormalization = function (t, e, n, r) {
        return (
          void 0 === t && (t = 5),
          void 0 === e && (e = 1),
          void 0 === n && (n = 1),
          void 0 === r && (r = 0.5),
          xt.localResponseNormalization(this, t, e, n, r)
        )
      }),
      (t.prototype.pool = function (t, e, n, r, o) {
        return this.throwIfDisposed(), xt.pool(this, t, e, n, r, o)
      }),
      (t.prototype.variable = function (t, e, n) {
        return (
          void 0 === t && (t = !0),
          this.throwIfDisposed(),
          yt().makeVariable(this, t, e, n)
        )
      }),
      (t.prototype.unsortedSegmentSum = function (t, e) {
        return this.throwIfDisposed(), xt.unsortedSegmentSum(this, t, e)
      }),
      (t.prototype.batchToSpaceND = function (t, e) {
        return this.throwIfDisposed(), xt.batchToSpaceND(this, t, e)
      }),
      (t.prototype.spaceToBatchND = function (t, e) {
        return this.throwIfDisposed(), xt.spaceToBatchND(this, t, e)
      }),
      (t.prototype.topk = function (t, e) {
        return (
          void 0 === t && (t = 1),
          void 0 === e && (e = !0),
          this.throwIfDisposed(),
          xt.topk(this, t, e)
        )
      }),
      (t.prototype.stridedSlice = function (t, e, n, r, o, a, i, s) {
        return (
          void 0 === r && (r = 0),
          void 0 === o && (o = 0),
          void 0 === a && (a = 0),
          void 0 === i && (i = 0),
          void 0 === s && (s = 0),
          this.throwIfDisposed(),
          xt.stridedSlice(this, t, e, n, r, o, a, i, s)
        )
      }),
      (t.prototype.depthToSpace = function (t, e) {
        return this.throwIfDisposed(), xt.depthToSpace(this, t, e)
      }),
      (t.prototype.fft = function () {
        return this.throwIfDisposed(), xt.spectral.fft(this)
      }),
      (t.prototype.ifft = function () {
        return this.throwIfDisposed(), xt.spectral.ifft(this)
      }),
      (t.prototype.rfft = function () {
        return this.throwIfDisposed(), xt.spectral.rfft(this)
      }),
      (t.prototype.irfft = function () {
        return this.throwIfDisposed(), xt.spectral.irfft(this)
      }),
      t
    )
  })()
  Object.defineProperty(wt, Symbol.hasInstance, {
    value: function (t) {
      return !!t && null != t.dataId && null != t.shape && null != t.dtype
    },
  })
  var Ct,
    Et,
    Rt,
    It,
    kt,
    St = (function (t) {
      function n(e, n, r, o) {
        var a = t.call(this, e.shape, e.dtype, e.dataId, o) || this
        return (a.trainable = n), (a.name = r), a
      }
      return (
        e(n, t),
        (n.prototype.assign = function (t) {
          if (t.dtype !== this.dtype)
            throw new Error(
              'dtype of the new value (' +
                t.dtype +
                ') and previous value (' +
                this.dtype +
                ') must match'
            )
          if (!S(t.shape, this.shape))
            throw new Error(
              'shape of the new value (' +
                t.shape +
                ') and previous value (' +
                this.shape +
                ') must match'
            )
          yt().disposeTensor(this),
            (this.dataId = t.dataId),
            yt().incRef(this, null)
        }),
        (n.prototype.dispose = function () {
          yt().disposeVariable(this), (this.isDisposedInternal = !0)
        }),
        n
      )
    })(wt)
  Object.defineProperty(St, Symbol.hasInstance, {
    value: function (t) {
      return t instanceof wt && null != t.assign && t.assign instanceof Function
    },
  }),
    (function (t) {
      ;(t.R0 = 'R0'),
        (t.R1 = 'R1'),
        (t.R2 = 'R2'),
        (t.R3 = 'R3'),
        (t.R4 = 'R4'),
        (t.R5 = 'R5'),
        (t.R6 = 'R6')
    })(Ct || (Ct = {})),
    (function (t) {
      ;(t.float32 = 'float32'),
        (t.int32 = 'int32'),
        (t.bool = 'int32'),
        (t.complex64 = 'complex64')
    })(Et || (Et = {})),
    (function (t) {
      ;(t.float32 = 'float32'),
        (t.int32 = 'int32'),
        (t.bool = 'bool'),
        (t.complex64 = 'complex64')
    })(Rt || (Rt = {})),
    (function (t) {
      ;(t.float32 = 'float32'),
        (t.int32 = 'float32'),
        (t.bool = 'float32'),
        (t.complex64 = 'complex64')
    })(It || (It = {})),
    (function (t) {
      ;(t.float32 = 'complex64'),
        (t.int32 = 'complex64'),
        (t.bool = 'complex64'),
        (t.complex64 = 'complex64')
    })(kt || (kt = {}))
  var At = { float32: It, int32: Et, bool: Rt, complex64: kt }
  function Dt(t, e) {
    if ('string' === t || 'string' === e) {
      if ('string' === t && 'string' === e) return 'string'
      throw new Error('Can not upcast ' + t + ' with ' + e)
    }
    return At[t][e]
  }
  function Tt(t) {
    return Dt(t, 'int32')
  }
  function Nt(t, e) {
    if (t.dtype === e.dtype) return [t, e]
    var n = Dt(t.dtype, e.dtype)
    return [t.cast(n), e.cast(n)]
  }
  function Ft(t, e) {
    C(t.dtype === e.dtype, function () {
      return (
        'The dtypes of the first(' +
        t.dtype +
        ') and second(' +
        e.dtype +
        ') input must match'
      )
    })
  }
  function _t(t) {
    var e = []
    return (
      (function t(e, n, r) {
        if (null == e) return
        if (e instanceof wt) return void n.push(e)
        if (((o = e), !Array.isArray(o) && 'object' != typeof o)) return
        var o
        var a = e
        for (var i in a) {
          var s = a[i]
          r.has(s) || (r.add(s), t(s, n, r))
        }
      })(t, e, new Set()),
      e
    )
  }
  var Ot,
    Mt = Object.freeze({
      makeTypesMatch: Nt,
      assertTypesMatch: Ft,
      isTensorInList: function (t, e) {
        return e.some(function (e) {
          return e.id === t.id
        })
      },
      getTensorsInContainer: _t,
    }),
    Bt = (function () {
      function t() {
        ;(this.registeredVariables = {}),
          (this.nextTapeNodeId = 0),
          (this.numBytes = 0),
          (this.numTensors = 0),
          (this.numStringTensors = 0),
          (this.numDataBuffers = 0),
          (this.gradientDepth = 0),
          (this.kernelDepth = 0),
          (this.scopeStack = []),
          (this.numDataMovesStack = []),
          (this.nextScopeId = 0),
          (this.tensorInfo = new WeakMap()),
          (this.profiling = !1),
          (this.activeProfile = {
            newBytes: 0,
            newTensors: 0,
            peakBytes: 0,
            kernels: [],
            result: null,
          })
      }
      return (
        (t.prototype.dispose = function () {
          for (var t in this.registeredVariables)
            this.registeredVariables[t].dispose()
        }),
        t
      )
    })(),
    Pt = (function () {
      function t(t) {
        ;(this.ENV = t),
          (this.registry = {}),
          (this.registryFactory = {}),
          (this.pendingBackendInitId = 0),
          (this.state = new Bt())
      }
      return (
        (t.prototype.ready = function () {
          return n(this, void 0, void 0, function () {
            var t, e, n
            return r(this, function (r) {
              switch (r.label) {
                case 0:
                  if (null != this.pendingBackendInit)
                    return [2, this.pendingBackendInit.then(function () {})]
                  if (null != this.backendInstance) return [2]
                  ;(t = this.getSortedBackends()), (e = 0), (r.label = 1)
                case 1:
                  return e < t.length
                    ? ((n = t[e]), [4, this.initializeBackend(n).success])
                    : [3, 5]
                case 2:
                  return r.sent() ? [4, this.setBackend(n)] : [3, 4]
                case 3:
                  return r.sent(), [2]
                case 4:
                  return e++, [3, 1]
                case 5:
                  throw new Error(
                    'Could not initialize any backends, all backend initializations failed.'
                  )
              }
            })
          })
        }),
        Object.defineProperty(t.prototype, 'backend', {
          get: function () {
            if (null != this.pendingBackendInit)
              throw new Error(
                "Backend '" +
                  this.backendName +
                  "' has not yet been initialized. Make sure to await tf.ready() or await tf.setBackend() before calling other methods"
              )
            if (null == this.backendInstance) {
              var t = this.initializeBackendsAndReturnBest(),
                e = t.name
              if (t.asyncInit)
                throw new Error(
                  "The highest priority backend '" +
                    e +
                    "' has not yet been initialized. Make sure to await tf.ready() or await tf.setBackend() before calling other methods"
                )
              this.setBackend(e)
            }
            return this.backendInstance
          },
          enumerable: !0,
          configurable: !0,
        }),
        (t.prototype.backendNames = function () {
          return Object.keys(this.registryFactory)
        }),
        (t.prototype.findBackend = function (t) {
          if (!(t in this.registry)) {
            if (!(t in this.registryFactory)) return null
            if (this.initializeBackend(t).asyncInit) return null
          }
          return this.registry[t]
        }),
        (t.prototype.findBackendFactory = function (t) {
          return t in this.registryFactory
            ? this.registryFactory[t].factory
            : null
        }),
        (t.prototype.registerBackend = function (t, e, n) {
          return (
            void 0 === n && (n = 1),
            t in this.registryFactory
              ? (console.warn(
                  t +
                    ' backend was already registered. Reusing existing backend factory.'
                ),
                !1)
              : ((this.registryFactory[t] = { factory: e, priority: n }), !0)
          )
        }),
        (t.prototype.setBackend = function (t) {
          return n(this, void 0, void 0, function () {
            var e, n, o
            return r(this, function (r) {
              switch (r.label) {
                case 0:
                  if (null == this.registryFactory[t])
                    throw new Error(
                      "Backend name '" + t + "' not found in registry"
                    )
                  return (
                    (this.backendName = t),
                    null != this.registry[t]
                      ? [3, 4]
                      : ((this.backendInstance = null),
                        (e = this.initializeBackend(t)),
                        (n = e.success),
                        e.asyncInit ? [4, n] : [3, 2])
                  )
                case 1:
                  return (o = r.sent()), [3, 3]
                case 2:
                  ;(o = n), (r.label = 3)
                case 3:
                  if (!o) return [2, !1]
                  r.label = 4
                case 4:
                  return (
                    (this.backendInstance = this.registry[t]),
                    this.setupRegisteredKernels(),
                    (this.profiler = new ut(this.backendInstance)),
                    [2, !0]
                  )
              }
            })
          })
        }),
        (t.prototype.setupRegisteredKernels = function () {
          var t = this
          f(this.backendName).forEach(function (e) {
            null != e.setupFunc && e.setupFunc(t.backendInstance)
          })
        }),
        (t.prototype.disposeRegisteredKernels = function (t) {
          var e = this
          f(t).forEach(function (n) {
            null != n.disposeFunc && n.disposeFunc(e.registry[t])
          })
        }),
        (t.prototype.initializeBackend = function (t) {
          var e = this,
            n = this.registryFactory[t]
          if (null == n)
            throw new Error(
              'Cannot initialize backend ' + t + ', no registration found.'
            )
          try {
            var r = n.factory()
            if (Promise.resolve(r) === r) {
              var o = ++this.pendingBackendInitId,
                a = r
                  .then(function (n) {
                    return (
                      !(o < e.pendingBackendInitId) &&
                      ((e.registry[t] = n), (e.pendingBackendInit = null), !0)
                    )
                  })
                  .catch(function (n) {
                    return (
                      !(o < e.pendingBackendInitId) &&
                      ((e.pendingBackendInit = null),
                      console.warn(
                        'Initialization of backend ' + t + ' failed'
                      ),
                      console.warn(n.stack || n.message),
                      !1)
                    )
                  })
              return (
                (this.pendingBackendInit = a), { success: a, asyncInit: !0 }
              )
            }
            return (this.registry[t] = r), { success: !0, asyncInit: !1 }
          } catch (e) {
            return (
              console.warn('Initialization of backend ' + t + ' failed'),
              console.warn(e.stack || e.message),
              { success: !1, asyncInit: !1 }
            )
          }
        }),
        (t.prototype.removeBackend = function (t) {
          if (!(t in this.registryFactory))
            throw new Error(t + ' backend not found in registry')
          this.backendName === t &&
            null != this.pendingBackendInit &&
            this.pendingBackendInitId++,
            t in this.registry &&
              (this.disposeRegisteredKernels(t),
              this.registry[t].dispose(),
              delete this.registry[t]),
            delete this.registryFactory[t],
            this.backendName === t &&
              ((this.pendingBackendInit = null),
              (this.backendName = null),
              (this.backendInstance = null))
        }),
        (t.prototype.getSortedBackends = function () {
          var t = this
          if (0 === Object.keys(this.registryFactory).length)
            throw new Error('No backend found in registry.')
          return Object.keys(this.registryFactory).sort(function (e, n) {
            return t.registryFactory[n].priority - t.registryFactory[e].priority
          })
        }),
        (t.prototype.initializeBackendsAndReturnBest = function () {
          for (var t = this.getSortedBackends(), e = 0; e < t.length; e++) {
            var n = t[e],
              r = this.initializeBackend(n),
              o = r.success,
              a = r.asyncInit
            if (a || o) return { name: n, asyncInit: a }
          }
          throw new Error(
            'Could not initialize any backends, all backend initializations failed.'
          )
        }),
        (t.prototype.moveData = function (t, e) {
          var n = this.state.tensorInfo.get(e),
            r = n.backend,
            o = this.readSync(e)
          r.disposeData(e),
            (n.backend = t),
            t.move(e, o, n.shape, n.dtype),
            this.shouldCheckForMemLeaks() &&
              this.state.numDataMovesStack[
                this.state.numDataMovesStack.length - 1
              ]++
        }),
        (t.prototype.tidy = function (t, e) {
          var n,
            r = this,
            o = null
          if (null == e) {
            if ('function' != typeof t)
              throw new Error('Please provide a function to tidy()')
            e = t
          } else {
            if ('string' != typeof t && !(t instanceof String))
              throw new Error(
                'When calling with two arguments, the first argument to tidy() must be a string'
              )
            if ('function' != typeof e)
              throw new Error(
                'When calling with two arguments, the 2nd argument to tidy() must be a function'
              )
            o = t
          }
          return this.scopedRun(
            function () {
              return r.startScope(o)
            },
            function () {
              return r.endScope(n)
            },
            function () {
              return (
                (n = e()) instanceof Promise &&
                  console.error('Cannot return a Promise inside of tidy.'),
                n
              )
            }
          )
        }),
        (t.prototype.scopedRun = function (t, e, n) {
          t()
          try {
            var r = n()
            return e(), r
          } catch (t) {
            throw (e(), t)
          }
        }),
        (t.prototype.nextTensorId = function () {
          return t.nextTensorId++
        }),
        (t.prototype.nextVariableId = function () {
          return t.nextVariableId++
        }),
        (t.prototype.clone = function (t) {
          var e = this.makeTensorFromDataId(t.dataId, t.shape, t.dtype),
            n = { x: t }
          return (
            this.addTapeNode(
              this.state.activeScope.name,
              n,
              [e],
              function (t) {
                return {
                  x: function () {
                    return t.toFloat()
                  },
                }
              },
              []
            ),
            e
          )
        }),
        (t.prototype.runKernel = function (t, e, n, r, o) {
          return this.runKernelFunc(null, e, null, t, n, r, o)
        }),
        (t.prototype.shouldCheckForMemLeaks = function () {
          return this.ENV.getBool('IS_TEST')
        }),
        (t.prototype.checkKernelForMemLeak = function (t, e, n) {
          var r = this.backend.numDataIds(),
            o = 0
          n.forEach(function (t) {
            o += 'complex64' === t.dtype ? 3 : 1
          })
          var a =
              this.state.numDataMovesStack[
                this.state.numDataMovesStack.length - 1
              ],
            i = r - e - o - a
          if (i > 0)
            throw new Error(
              "Backend '" +
                this.backendName +
                "' has an internal memory leak (" +
                i +
                " data ids) after running '" +
                t +
                "'"
            )
        }),
        (t.prototype.runKernelFunc = function (t, e, n, r, o, a, i) {
          var s,
            u = this
          void 0 === a && (a = []), void 0 === i && (i = [])
          var c = [],
            h = this.isTapeOn()
          null == r &&
            (r =
              null != this.state.activeScope ? this.state.activeScope.name : '')
          var f,
            d = function (t) {
              h &&
                (c = t.map(function (t) {
                  return u.keep(u.clone(t))
                }))
            },
            p = this.state.numBytes,
            v = this.state.numTensors
          this.shouldCheckForMemLeaks() && this.state.numDataMovesStack.push(0)
          var m,
            g = l(r, this.backendName)
          return (
            (f =
              null != g
                ? function () {
                    var t = u.backend.numDataIds()
                    m = g.kernelFunc({
                      inputs: e,
                      attrs: o,
                      backend: u.backend,
                    })
                    var n = Array.isArray(m) ? m : [m]
                    u.shouldCheckForMemLeaks() &&
                      u.checkKernelForMemLeak(r, t, n)
                    var s = n.map(function (t) {
                        var e = t.dataId,
                          n = t.shape,
                          r = t.dtype
                        return u.makeTensorFromDataId(e, n, r)
                      }),
                      c = s.filter(function (t, e) {
                        return i[e]
                      })
                    return d((a || []).slice().concat(c)), s
                  }
                : function () {
                    var e = u.backend.numDataIds()
                    m = u.tidy(function () {
                      return t(u.backend, d)
                    })
                    var n = Array.isArray(m) ? m : [m]
                    return (
                      u.shouldCheckForMemLeaks() &&
                        u.checkKernelForMemLeak(r, e, n),
                      n
                    )
                  }),
            this.scopedRun(
              function () {
                return u.state.kernelDepth++
              },
              function () {
                return u.state.kernelDepth--
              },
              function () {
                s = u.ENV.getBool('DEBUG')
                  ? u.profiler.profileKernel(r, e, function () {
                      return f()
                    })
                  : f()
              }
            ),
            h && this.addTapeNode(r, e, s, n, c),
            this.state.profiling &&
              this.state.activeProfile.kernels.push({
                name: r,
                bytesAdded: this.state.numBytes - p,
                totalBytesSnapshot: this.state.numBytes,
                tensorsAdded: this.state.numTensors - v,
                totalTensorsSnapshot: this.state.numTensors,
                inputShapes: Object.keys(e).map(function (t) {
                  return e[t].shape
                }),
                outputShapes: s.map(function (t) {
                  return t.shape
                }),
              }),
            Array.isArray(m) ? s : s[0]
          )
        }),
        (t.prototype.makeTensor = function (t, e, n, r) {
          if (null == t)
            throw new Error('Values passed to engine.makeTensor() are null')
          ;(n = n || 'float32'), (r = r || this.backend)
          var o = t
          'string' === n &&
            H(t[0]) &&
            (o = t.map(function (t) {
              return rt(t)
            }))
          var a = r.write(o, e, n),
            i = new wt(e, n, a, this.nextTensorId())
          if ((this.incRef(i, r), 'string' === n)) {
            var s = this.state.tensorInfo.get(a),
              u = G(o)
            ;(this.state.numBytes += u - s.bytes), (s.bytes = u)
          }
          return i
        }),
        (t.prototype.makeTensorFromDataId = function (t, e, n, r) {
          var o = new wt(e, (n = n || 'float32'), t, this.nextTensorId())
          return this.incRef(o, r), o
        }),
        (t.prototype.makeVariable = function (t, e, n, r) {
          void 0 === e && (e = !0),
            (n = n || this.nextVariableId().toString()),
            null != r && r !== t.dtype && (t = t.asType(r))
          var o = new St(t, e, n, this.nextTensorId())
          if (null != this.state.registeredVariables[o.name])
            throw new Error(
              'Variable with name ' + o.name + ' was already registered'
            )
          return (
            (this.state.registeredVariables[o.name] = o),
            this.incRef(o, this.backend),
            o
          )
        }),
        (t.prototype.incRef = function (t, e) {
          var n = this.state.tensorInfo.has(t.dataId)
            ? this.state.tensorInfo.get(t.dataId).refCount
            : 0
          if (
            (this.state.numTensors++,
            'string' === t.dtype && this.state.numStringTensors++,
            0 === n)
          ) {
            this.state.numDataBuffers++
            var r = 0
            'complex64' !== t.dtype &&
              'string' !== t.dtype &&
              (r = t.size * z(t.dtype)),
              this.state.tensorInfo.set(t.dataId, {
                backend: e || this.backend,
                dtype: t.dtype,
                shape: t.shape,
                bytes: r,
                refCount: 0,
              }),
              (this.state.numBytes += r)
          }
          this.state.tensorInfo.get(t.dataId).refCount++,
            t instanceof St || this.track(t)
        }),
        (t.prototype.disposeTensor = function (t) {
          if (this.state.tensorInfo.has(t.dataId)) {
            this.state.numTensors--,
              'string' === t.dtype && this.state.numStringTensors--
            var e = this.state.tensorInfo.get(t.dataId)
            e.refCount <= 1
              ? ('complex64' !== t.dtype && (this.state.numBytes -= e.bytes),
                this.state.numDataBuffers--,
                e.backend.disposeData(t.dataId),
                this.state.tensorInfo.delete(t.dataId))
              : this.state.tensorInfo.get(t.dataId).refCount--
          }
        }),
        (t.prototype.disposeVariables = function () {
          for (var t in this.state.registeredVariables) {
            var e = this.state.registeredVariables[t]
            this.disposeVariable(e)
          }
        }),
        (t.prototype.disposeVariable = function (t) {
          this.disposeTensor(t),
            null != this.state.registeredVariables[t.name] &&
              delete this.state.registeredVariables[t.name]
        }),
        (t.prototype.memory = function () {
          var t = this.backend.memory()
          return (
            (t.numTensors = this.state.numTensors),
            (t.numDataBuffers = this.state.numDataBuffers),
            (t.numBytes = this.state.numBytes),
            this.state.numStringTensors > 0 &&
              ((t.unreliable = !0),
              null == t.reasons && (t.reasons = []),
              t.reasons.push(
                'Memory usage by string tensors is approximate (2 bytes per character)'
              )),
            t
          )
        }),
        (t.prototype.profile = function (t) {
          return n(this, void 0, void 0, function () {
            var e, n
            return r(this, function (r) {
              return (
                (this.state.profiling = !0),
                (e = this.state.numBytes),
                (n = this.state.numTensors),
                (this.state.activeProfile.kernels = []),
                (this.state.activeProfile.result = t()),
                (this.state.profiling = !1),
                (this.state.activeProfile.peakBytes = Math.max.apply(
                  Math,
                  this.state.activeProfile.kernels.map(function (t) {
                    return t.totalBytesSnapshot
                  })
                )),
                (this.state.activeProfile.newBytes = this.state.numBytes - e),
                (this.state.activeProfile.newTensors =
                  this.state.numTensors - n),
                [2, this.state.activeProfile]
              )
            })
          })
        }),
        (t.prototype.isTapeOn = function () {
          return this.state.gradientDepth > 0 && 0 === this.state.kernelDepth
        }),
        (t.prototype.addTapeNode = function (t, e, n, r, o) {
          var a = this,
            i = {
              id: this.state.nextTapeNodeId++,
              kernelName: t,
              inputs: e,
              outputs: n,
              saved: o,
            },
            s = h(t)
          null != s && (r = s.gradFunc),
            null != r &&
              (i.gradient = function (t) {
                return (
                  (t = t.map(function (t, e) {
                    if (null == t) {
                      var r = n[e],
                        o = tt(r.size, r.dtype)
                      return a.makeTensor(o, r.shape, r.dtype)
                    }
                    return t
                  })),
                  r(t.length > 1 ? t : t[0], o)
                )
              }),
            this.state.activeTape.push(i)
        }),
        (t.prototype.keep = function (t) {
          return (t.kept = !0), t
        }),
        (t.prototype.startTape = function () {
          0 === this.state.gradientDepth && (this.state.activeTape = []),
            this.state.gradientDepth++
        }),
        (t.prototype.endTape = function () {
          this.state.gradientDepth--
        }),
        (t.prototype.startScope = function (t) {
          var e = {
            track: [],
            name: 'unnamed scope',
            id: this.state.nextScopeId++,
          }
          t && (e.name = t),
            this.state.scopeStack.push(e),
            (this.state.activeScope = e)
        }),
        (t.prototype.endScope = function (t) {
          for (
            var e = this,
              n = _t(t),
              r = new Set(
                n.map(function (t) {
                  return t.id
                })
              ),
              o = 0;
            o < this.state.activeScope.track.length;
            o++
          ) {
            var a = this.state.activeScope.track[o]
            a.kept || r.has(a.id) || a.dispose()
          }
          var i = this.state.scopeStack.pop()
          ;(this.state.activeScope =
            0 === this.state.scopeStack.length
              ? null
              : this.state.scopeStack[this.state.scopeStack.length - 1]),
            n.forEach(function (t) {
              t.kept || t.scopeId !== i.id || e.track(t)
            })
        }),
        (t.prototype.gradients = function (t, e, n, r) {
          var o = this
          if (
            (void 0 === r && (r = !1),
            C(e.length > 0, function () {
              return 'gradients() received an empty list of xs.'
            }),
            null != n && 'float32' !== n.dtype)
          )
            throw new Error(
              "dy must have 'float32' dtype, but has '" + n.dtype + "'"
            )
          var a = this.scopedRun(
            function () {
              return o.startTape()
            },
            function () {
              return o.endTape()
            },
            function () {
              return o.tidy('forward', t)
            }
          )
          C(a instanceof wt, function () {
            return 'The result y returned by f() must be a tensor.'
          })
          var i = (function (t, e, n) {
            for (var r = {}, o = {}, a = 0; a < e.length; a++) r[e[a].id] = !0
            for (a = 0; a < t.length; a++) {
              var i = (p = t[a]).inputs
              for (var s in i) {
                for (var u = i[s], c = !1, l = 0; l < e.length; l++)
                  if (r[u.id]) {
                    p.outputs.forEach(function (t) {
                      return (r[t.id] = !0)
                    }),
                      (c = !0),
                      (o[p.id] = !0)
                    break
                  }
                if (c) break
              }
            }
            var h = {}
            h[n.id] = !0
            var f = {}
            for (a = t.length - 1; a >= 0; a--)
              for (i = (p = t[a]).inputs, l = 0; l < p.outputs.length; l++)
                if (h[p.outputs[l].id]) {
                  for (var s in i) (h[i[s].id] = !0), (f[p.id] = !0)
                  break
                }
            var d = []
            for (a = 0; a < t.length; a++) {
              var p
              if (o[(p = t[a]).id] && f[p.id]) {
                var v = {}
                for (var s in p.inputs) {
                  var m = p.inputs[s]
                  r[m.id] && (v[s] = m)
                }
                var g = Object.assign({}, p)
                ;(g.inputs = v), (g.outputs = p.outputs), d.push(g)
              }
            }
            return d
          })(this.state.activeTape, e, a)
          if (!r && 0 === i.length && e.length > 0)
            throw new Error(
              'Cannot compute gradient of y=f(x) with respect to x. Make sure that the f you passed encloses all operations that lead from x to y.'
            )
          return this.tidy('backward', function () {
            var t,
              r,
              s = {}
            ;(s[a.id] =
              null == n
                ? ((t = a.shape),
                  (r = Z(k(t), 'float32')),
                  Lt.makeTensor(r, t, 'float32'))
                : n),
              (function (t, e, n) {
                for (
                  var r = function (r) {
                      var o = e[r],
                        a = []
                      if (
                        (o.outputs.forEach(function (e) {
                          var n = t[e.id]
                          null != n ? a.push(n) : a.push(null)
                        }),
                        null == o.gradient)
                      )
                        throw new Error(
                          'Cannot compute gradient: gradient function not found for ' +
                            o.kernelName +
                            '.'
                        )
                      var i = o.gradient(a),
                        s = function (e) {
                          if (!(e in i))
                            throw new Error(
                              'Cannot backprop through input ' +
                                e +
                                '. Available gradients found: ' +
                                Object.keys(i) +
                                '.'
                            )
                          var r = n(function () {
                            return i[e]()
                          })
                          if ('float32' !== r.dtype)
                            throw new Error(
                              'Error in gradient for op ' +
                                o.kernelName +
                                '. The gradient of input ' +
                                e +
                                " must have 'float32' dtype, but has '" +
                                r.dtype +
                                "'"
                            )
                          var a = o.inputs[e]
                          if (!S(r.shape, a.shape))
                            throw new Error(
                              'Error in gradient for op ' +
                                o.kernelName +
                                ". The gradient of input '" +
                                e +
                                "' has shape '" +
                                r.shape +
                                "', which does not match the shape of the input '" +
                                a.shape +
                                "'"
                            )
                          if (null == t[a.id]) t[a.id] = r
                          else {
                            var s = t[a.id]
                            ;(t[a.id] = s.add(r)), s.dispose()
                          }
                        }
                      for (var u in o.inputs) s(u)
                    },
                    o = e.length - 1;
                  o >= 0;
                  o--
                )
                  r(o)
              })(s, i, function (t) {
                return o.tidy(t)
              })
            var u = e.map(function (t) {
              return s[t.id]
            })
            return (
              0 === o.state.gradientDepth &&
                (o.state.activeTape.forEach(function (t) {
                  for (var e = 0, n = t.saved; e < n.length; e++) {
                    n[e].dispose()
                  }
                }),
                (o.state.activeTape = null)),
              { value: a, grads: u }
            )
          })
        }),
        (t.prototype.customGrad = function (t) {
          var e = this
          return (
            C(X(t), function () {
              return 'The f passed in customGrad(f) must be a function.'
            }),
            function () {
              for (var n, r = [], o = 0; o < arguments.length; o++)
                r[o] = arguments[o]
              C(
                r.every(function (t) {
                  return t instanceof wt
                }),
                function () {
                  return 'The args passed in customGrad(f)(x1, x2,...) must all be tensors'
                }
              )
              var a = {}
              return (
                r.forEach(function (t, e) {
                  a[e] = t
                }),
                e.runKernelFunc(
                  function (e, o) {
                    return (
                      C(
                        (n = t.apply(void 0, r.concat([o]))).value instanceof
                          wt,
                        function () {
                          return 'The function f passed in customGrad(f) must return an object where `obj.value` is a tensor'
                        }
                      ),
                      C(X(n.gradFunc), function () {
                        return 'The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function.'
                      }),
                      n.value
                    )
                  },
                  a,
                  function (t, e) {
                    var o = n.gradFunc(t, e),
                      a = Array.isArray(o) ? o : [o]
                    C(a.length === r.length, function () {
                      return 'The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function that returns the same number of tensors as inputs passed to f(...).'
                    }),
                      C(
                        a.every(function (t) {
                          return t instanceof wt
                        }),
                        function () {
                          return 'The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function that returns a list of only tensors.'
                        }
                      )
                    var i = {}
                    return (
                      a.forEach(function (t, e) {
                        i[e] = function () {
                          return t
                        }
                      }),
                      i
                    )
                  }
                )
              )
            }
          )
        }),
        (t.prototype.readSync = function (t) {
          return this.state.tensorInfo.get(t).backend.readSync(t)
        }),
        (t.prototype.read = function (t) {
          return this.state.tensorInfo.get(t).backend.read(t)
        }),
        (t.prototype.time = function (t) {
          return n(this, void 0, void 0, function () {
            var e, n
            return r(this, function (r) {
              switch (r.label) {
                case 0:
                  return (e = et()), [4, this.backend.time(t)]
                case 1:
                  return ((n = r.sent()).wallMs = et() - e), [2, n]
              }
            })
          })
        }),
        (t.prototype.track = function (t) {
          return (
            null != this.state.activeScope &&
              ((t.scopeId = this.state.activeScope.id),
              this.state.activeScope.track.push(t)),
            t
          )
        }),
        Object.defineProperty(t.prototype, 'registeredVariables', {
          get: function () {
            return this.state.registeredVariables
          },
          enumerable: !0,
          configurable: !0,
        }),
        (t.prototype.reset = function () {
          for (var t in (this.pendingBackendInitId++,
          this.state.dispose(),
          this.ENV.reset(),
          (this.state = new Bt()),
          this.registry))
            this.disposeRegisteredKernels(t),
              this.registry[t].dispose(),
              delete this.registry[t]
          ;(this.backendName = null),
            (this.backendInstance = null),
            (this.pendingBackendInit = null)
        }),
        (t.nextTensorId = 0),
        (t.nextVariableId = 0),
        t
      )
    })()
  var Lt = (function () {
    var t = (function () {
      if (null == Ot) {
        var t = void 0
        if ('undefined' != typeof window) t = window
        else if ('undefined' != typeof global) t = global
        else if ('undefined' != typeof process) t = process
        else {
          if ('undefined' == typeof self)
            throw new Error('Could not find a global object')
          t = self
        }
        Ot = t
      }
      return Ot
    })()
    if (null == t._tfengine) {
      var e = new o(t)
      t._tfengine = new Pt(e)
    }
    return (
      (function (t) {
        s = t
      })(t._tfengine.ENV),
      (yt = function () {
        return t._tfengine
      }),
      t._tfengine
    )
  })()
  function Wt() {
    return (
      ('undefined' != typeof window && null != window.document) ||
      'undefined' != typeof WorkerGlobalScope
    )
  }
  var Ut = i()
  Ut.registerFlag(
    'DEBUG',
    function () {
      return !1
    },
    function (t) {
      t &&
        console.warn(
          'Debugging mode is ON. The output of every math call will be downloaded to CPU and checked for NaNs. This significantly impacts performance.'
        )
    }
  ),
    Ut.registerFlag('IS_BROWSER', function () {
      return Wt()
    }),
    Ut.registerFlag('IS_NODE', function () {
      return (
        'undefined' != typeof process &&
        void 0 !== process.versions &&
        void 0 !== process.versions.node
      )
    }),
    Ut.registerFlag('IS_CHROME', function () {
      return (
        'undefined' != typeof navigator &&
        null != navigator &&
        null != navigator.userAgent &&
        /Chrome/.test(navigator.userAgent) &&
        /Google Inc/.test(navigator.vendor)
      )
    }),
    Ut.registerFlag('PROD', function () {
      return !1
    }),
    Ut.registerFlag('TENSORLIKE_CHECK_SHAPE_CONSISTENCY', function () {
      return Ut.getBool('DEBUG')
    }),
    Ut.registerFlag('DEPRECATION_WARNINGS_ENABLED', function () {
      return !0
    }),
    Ut.registerFlag('IS_TEST', function () {
      return !1
    })
  var Vt,
    zt,
    Gt,
    Ht = {},
    qt = {
      alpha: !1,
      antialias: !1,
      premultipliedAlpha: !1,
      preserveDrawingBuffer: !1,
      depth: !1,
      stencil: !1,
      failIfMajorPerformanceCaveat: !0,
    }
  function Kt(t, e) {
    Ht[t] = e
  }
  function jt(t) {
    t in Ht ||
      (Ht[t] = (function (t) {
        if (1 !== t && 2 !== t)
          throw new Error(
            'Cannot get WebGL rendering context, WebGL is disabled.'
          )
        var e = (function (t) {
          if ('undefined' != typeof OffscreenCanvas && 2 === t)
            return new OffscreenCanvas(300, 150)
          if ('undefined' != typeof document)
            return document.createElement('canvas')
          throw new Error('Cannot create a canvas in this context')
        })(t)
        if (
          (e.addEventListener(
            'webglcontextlost',
            function (e) {
              e.preventDefault(), delete Ht[t]
            },
            !1
          ),
          1 === t)
        )
          return (
            e.getContext('webgl', qt) || e.getContext('experimental-webgl', qt)
          )
        return e.getContext('webgl2', qt)
      })(t))
    var e = Ht[t]
    return e.isContextLost()
      ? (delete Ht[t], jt(t))
      : (e.disable(e.DEPTH_TEST),
        e.disable(e.STENCIL_TEST),
        e.disable(e.BLEND),
        e.disable(e.DITHER),
        e.disable(e.POLYGON_OFFSET_FILL),
        e.disable(e.SAMPLE_COVERAGE),
        e.enable(e.SCISSOR_TEST),
        e.enable(e.CULL_FACE),
        e.cullFace(e.BACK),
        Ht[t])
  }
  function Xt(t, e) {
    return [e, t]
  }
  function Yt(t) {
    var e = k(t)
    return T(Math.ceil(e / 4))
  }
  function $t(t, e) {
    return [Math.max(1, Math.ceil(e / 2)), Math.max(1, Math.ceil(t / 2))]
  }
  function Qt(t, e) {
    var n,
      r,
      o,
      a,
      s,
      u,
      c,
      l,
      h,
      f = t
    return (
      2 === i().getNumber('WEBGL_VERSION')
        ? ((n = f.R32F),
          (r = f.R16F),
          (o = f.RGBA16F),
          (a = f.RGBA32F),
          (s = f.RED),
          (u = 4),
          (c = 1),
          (l = f.HALF_FLOAT),
          (h = f.FLOAT))
        : ((n = t.RGBA),
          (r = t.RGBA),
          (o = t.RGBA),
          (a = f.RGBA),
          (s = t.RGBA),
          (u = 4),
          (c = 4),
          (l = null != e ? e.HALF_FLOAT_OES : null),
          (h = t.FLOAT)),
      {
        internalFormatFloat: n,
        internalFormatHalfFloat: r,
        internalFormatPackedHalfFloat: o,
        internalFormatPackedFloat: a,
        textureFormatFloat: s,
        downloadTextureFormat: t.RGBA,
        downloadUnpackNumChannels: u,
        defaultNumChannels: c,
        textureTypeHalfFloat: l,
        textureTypeFloat: h,
      }
    )
  }
  function Jt(t, e, n) {
    var r = n()
    return (
      e &&
        (function (t) {
          var e = t.getError()
          if (e !== t.NO_ERROR) throw new Error('WebGL Error: ' + ne(t, e))
        })(t),
      r
    )
  }
  !(function (t) {
    ;(t[(t.DENSE = 0)] = 'DENSE'), (t[(t.SHARED_BATCH = 1)] = 'SHARED_BATCH')
  })(Vt || (Vt = {})),
    (function (t) {
      ;(t[(t.RENDER = 0)] = 'RENDER'),
        (t[(t.UPLOAD = 1)] = 'UPLOAD'),
        (t[(t.PIXELS = 2)] = 'PIXELS'),
        (t[(t.DOWNLOAD = 3)] = 'DOWNLOAD')
    })(zt || (zt = {})),
    (function (t) {
      ;(t[(t.UNPACKED_FLOAT16 = 0)] = 'UNPACKED_FLOAT16'),
        (t[(t.UNPACKED_FLOAT32 = 1)] = 'UNPACKED_FLOAT32'),
        (t[(t.PACKED_4X1_UNSIGNED_BYTE = 2)] = 'PACKED_4X1_UNSIGNED_BYTE'),
        (t[(t.PACKED_2X2_FLOAT32 = 3)] = 'PACKED_2X2_FLOAT32'),
        (t[(t.PACKED_2X2_FLOAT16 = 4)] = 'PACKED_2X2_FLOAT16')
    })(Gt || (Gt = {}))
  var Zt = 5.96e-8,
    te = 65504
  function ee(t) {
    return !!(
      i().getBool('WEBGL_RENDER_FLOAT32_ENABLED') ||
      0 === t ||
      (Zt < Math.abs(t) && Math.abs(t) < te)
    )
  }
  function ne(t, e) {
    switch (e) {
      case t.NO_ERROR:
        return 'NO_ERROR'
      case t.INVALID_ENUM:
        return 'INVALID_ENUM'
      case t.INVALID_VALUE:
        return 'INVALID_VALUE'
      case t.INVALID_OPERATION:
        return 'INVALID_OPERATION'
      case t.INVALID_FRAMEBUFFER_OPERATION:
        return 'INVALID_FRAMEBUFFER_OPERATION'
      case t.OUT_OF_MEMORY:
        return 'OUT_OF_MEMORY'
      case t.CONTEXT_LOST_WEBGL:
        return 'CONTEXT_LOST_WEBGL'
      default:
        return 'Unknown error code ' + e
    }
  }
  function re(t, e, n) {
    return ke(
      t,
      e,
      function () {
        return t.getExtension(n)
      },
      'Extension "' + n + '" not supported on this browser.'
    )
  }
  function oe(t, e, n) {
    var r = ke(
      t,
      e,
      function () {
        return t.createShader(t.VERTEX_SHADER)
      },
      'Unable to create vertex WebGLShader.'
    )
    if (
      (Jt(t, e, function () {
        return t.shaderSource(r, n)
      }),
      Jt(t, e, function () {
        return t.compileShader(r)
      }),
      !1 === t.getShaderParameter(r, t.COMPILE_STATUS))
    )
      throw (
        (console.log(t.getShaderInfoLog(r)),
        new Error('Failed to compile vertex shader.'))
      )
    return r
  }
  function ae(t, e, n) {
    var r = ke(
      t,
      e,
      function () {
        return t.createShader(t.FRAGMENT_SHADER)
      },
      'Unable to create fragment WebGLShader.'
    )
    if (
      (Jt(t, e, function () {
        return t.shaderSource(r, n)
      }),
      Jt(t, e, function () {
        return t.compileShader(r)
      }),
      !1 === t.getShaderParameter(r, t.COMPILE_STATUS))
    )
      throw (
        ((function (t, e) {
          var n = ue.exec(e)
          if (null == n)
            return (
              console.log("Couldn't parse line number in error: " + e),
              void console.log(t)
            )
          for (
            var r = +n[1],
              o = t.split('\n'),
              a = o.length.toString().length + 2,
              i = o.map(function (t, e) {
                return N((e + 1).toString(), a) + t
              }),
              s = 0,
              u = 0;
            u < i.length;
            u++
          )
            s = Math.max(i[u].length, s)
          var c = i.slice(0, r - 1),
            l = i.slice(r - 1, r),
            h = i.slice(r)
          console.log(c.join('\n')),
            console.log(e.split('\n')[0]),
            console.log(
              '%c ' + N(l[0], s),
              'border:1px solid red; background-color:#e3d2d2; color:#a61717'
            ),
            console.log(h.join('\n'))
        })(n, t.getShaderInfoLog(r)),
        new Error('Failed to compile fragment shader.'))
      )
    return r
  }
  var ie,
    se,
    ue = /ERROR: [0-9]+:([0-9]+):/g
  function ce(t, e) {
    return ke(
      t,
      e,
      function () {
        return t.createProgram()
      },
      'Unable to create WebGLProgram.'
    )
  }
  function le(t, e, n) {
    if (
      (Jt(t, e, function () {
        return t.linkProgram(n)
      }),
      !1 === t.getProgramParameter(n, t.LINK_STATUS))
    )
      throw (
        (console.log(t.getProgramInfoLog(n)),
        new Error('Failed to link vertex and fragment shaders.'))
      )
  }
  function he(t, e, n) {
    if (
      (Jt(t, e, function () {
        return t.validateProgram(n)
      }),
      !1 === t.getProgramParameter(n, t.VALIDATE_STATUS))
    )
      throw (
        (console.log(t.getProgramInfoLog(n)),
        new Error('Shader program validation failed.'))
      )
  }
  function fe(t, e, n) {
    var r = ke(
      t,
      e,
      function () {
        return t.createBuffer()
      },
      'Unable to create WebGLBuffer'
    )
    return (
      Jt(t, e, function () {
        return t.bindBuffer(t.ARRAY_BUFFER, r)
      }),
      Jt(t, e, function () {
        return t.bufferData(t.ARRAY_BUFFER, n, t.STATIC_DRAW)
      }),
      r
    )
  }
  function de(t, e, n) {
    var r = ke(
      t,
      e,
      function () {
        return t.createBuffer()
      },
      'Unable to create WebGLBuffer'
    )
    return (
      Jt(t, e, function () {
        return t.bindBuffer(t.ELEMENT_ARRAY_BUFFER, r)
      }),
      Jt(t, e, function () {
        return t.bufferData(t.ELEMENT_ARRAY_BUFFER, n, t.STATIC_DRAW)
      }),
      r
    )
  }
  function pe(t, e) {
    return ke(
      t,
      e,
      function () {
        return t.createTexture()
      },
      'Unable to create WebGLTexture.'
    )
  }
  function ve(t, e) {
    var n = i().getNumber('WEBGL_MAX_TEXTURE_SIZE')
    if (t <= 0 || e <= 0) {
      var r = '[' + t + 'x' + e + ']'
      throw new Error('Requested texture size ' + r + ' is invalid.')
    }
    if (t > n || e > n) {
      r = '[' + t + 'x' + e + ']'
      throw new Error(
        'Requested texture size ' +
          r +
          ' greater than WebGL maximum on this browser / GPU ' +
          ('[' + n + 'x' + n + ']') +
          '.'
      )
    }
  }
  function me(t, e) {
    return ke(
      t,
      e,
      function () {
        return t.createFramebuffer()
      },
      'Unable to create WebGLFramebuffer.'
    )
  }
  function ge(t, e, n, r, o, a, i, s) {
    var u = t.getAttribLocation(n, r)
    return (
      -1 !== u &&
      (Jt(t, e, function () {
        return t.bindBuffer(t.ARRAY_BUFFER, o)
      }),
      Jt(t, e, function () {
        return t.vertexAttribPointer(u, a, t.FLOAT, !1, i, s)
      }),
      Jt(t, e, function () {
        return t.enableVertexAttribArray(u)
      }),
      !0)
    )
  }
  function ye(t, e, n, r) {
    Se(t, r),
      Jt(t, e, function () {
        return t.activeTexture(t.TEXTURE0 + r)
      }),
      Jt(t, e, function () {
        return t.bindTexture(t.TEXTURE_2D, n)
      })
  }
  function xe(t, e, n, r) {
    return ke(
      t,
      e,
      function () {
        return t.getUniformLocation(n, r)
      },
      'uniform "' + r + '" not present in program.'
    )
  }
  function be(t, e, n) {
    return t.getUniformLocation(e, n)
  }
  function we(t, e, n, r, o, a) {
    Jt(t, e, function () {
      return ye(t, e, r, a)
    }),
      Jt(t, e, function () {
        return t.uniform1i(o, a)
      })
  }
  function Ce(t, e, n, r) {
    Jt(t, e, function () {
      return t.bindFramebuffer(t.FRAMEBUFFER, r)
    }),
      Jt(t, e, function () {
        return t.framebufferTexture2D(
          t.FRAMEBUFFER,
          t.COLOR_ATTACHMENT0,
          t.TEXTURE_2D,
          n,
          0
        )
      })
  }
  function Ee(t, e, n) {
    Jt(t, e, function () {
      return t.bindFramebuffer(t.FRAMEBUFFER, n)
    }),
      Jt(t, e, function () {
        return t.framebufferTexture2D(
          t.FRAMEBUFFER,
          t.COLOR_ATTACHMENT0,
          t.TEXTURE_2D,
          null,
          0
        )
      })
  }
  function Re(t) {
    var e = t.checkFramebufferStatus(t.FRAMEBUFFER)
    if (e !== t.FRAMEBUFFER_COMPLETE)
      throw new Error('Error binding framebuffer: ' + Ie(t, e))
  }
  function Ie(t, e) {
    switch (e) {
      case t.FRAMEBUFFER_INCOMPLETE_ATTACHMENT:
        return 'FRAMEBUFFER_INCOMPLETE_ATTACHMENT'
      case t.FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT:
        return 'FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT'
      case t.FRAMEBUFFER_INCOMPLETE_DIMENSIONS:
        return 'FRAMEBUFFER_INCOMPLETE_DIMENSIONS'
      case t.FRAMEBUFFER_UNSUPPORTED:
        return 'FRAMEBUFFER_UNSUPPORTED'
      default:
        return 'unknown error ' + e
    }
  }
  function ke(t, e, n, r) {
    var o = Jt(t, e, function () {
      return n()
    })
    if (null == o) throw new Error(r)
    return o
  }
  function Se(t, e) {
    var n = t.MAX_COMBINED_TEXTURE_IMAGE_UNITS - 1,
      r = e + t.TEXTURE0
    if (r < t.TEXTURE0 || r > n)
      throw new Error(
        'textureUnit must be in ' + ('[gl.TEXTURE0, gl.TEXTURE' + n + ']') + '.'
      )
  }
  function Ae(t, e) {
    return void 0 === e && (e = 2), k(t.slice(0, t.length - e))
  }
  function De(t) {
    if (0 === t.length)
      throw Error('Cannot get rows and columns of an empty shape array.')
    return [t.length > 1 ? t[t.length - 2] : 1, t[t.length - 1]]
  }
  function Te(t) {
    var e = [1, 1, 1]
    return (
      0 === t.length ||
        (1 === t.length && 1 === t[0]) ||
        (e = [Ae(t)].concat(De(t))),
      e
    )
  }
  function Ne(t, e) {
    var n
    void 0 === e && (e = !1)
    var r = i().getNumber('WEBGL_MAX_TEXTURE_SIZE')
    if (
      (e &&
        ((r *= 2),
        1 ===
          (t = t.map(function (e, n) {
            return n >= t.length - 2 ? b(t[n]) : t[n]
          })).length && (t = [2, t[0]])),
      2 !== t.length)
    ) {
      var o = M(t)
      t = o.newShape
    }
    var a = k(t)
    if (t.length <= 1 && a <= r) return [1, a]
    if (2 === t.length && t[0] <= r && t[1] <= r) return t
    if (3 === t.length && t[0] * t[1] <= r && t[2] <= r)
      return [t[0] * t[1], t[2]]
    if (3 === t.length && t[0] <= r && t[1] * t[2] <= r)
      return [t[0], t[1] * t[2]]
    if (4 === t.length && t[0] * t[1] * t[2] <= r && t[3] <= r)
      return [t[0] * t[1] * t[2], t[3]]
    if (4 === t.length && t[0] <= r && t[1] * t[2] * t[3] <= r)
      return [t[0], t[1] * t[2] * t[3]]
    if (e) {
      var s = Ae(t),
        u = 2,
        c = 2
      return (
        t.length && ((u = (n = De(t))[0]), (c = n[1])),
        T((a = s * (u / 2) * (c / 2))).map(function (t) {
          return 2 * t
        })
      )
    }
    return T(a)
  }
  function Fe(t) {
    return t % 2 == 0
  }
  function _e(t, e) {
    if (S((t = t.slice(-2)), (e = e.slice(-2)))) return !0
    if (!t.length || !e.length) return !0
    if (0 === t[0] || 0 === t[1] || 0 === e[0] || 0 === e[1]) return !0
    if (t.length !== e.length) {
      var n = t.slice(-1)[0],
        r = e.slice(-1)[0]
      if (n === r) return !0
      if (Fe(n) && Fe(r) && (1 === t[0] || 1 === e[0])) return !0
    }
    return t[1] === e[1] && Fe(t[0]) && Fe(e[0])
  }
  function Oe(t) {
    if (null == ie) {
      var e = jt(t)
      ie = e.getParameter(e.MAX_TEXTURE_SIZE)
    }
    return ie
  }
  function Me(t) {
    if (null == se) {
      var e = jt(t)
      se = e.getParameter(e.MAX_TEXTURE_IMAGE_UNITS)
    }
    return Math.min(16, se)
  }
  function Be(t) {
    if (0 === t) return 0
    var e = jt(t)
    return Pe(e, 'EXT_disjoint_timer_query_webgl2') && 2 === t
      ? 2
      : Pe(e, 'EXT_disjoint_timer_query')
      ? 1
      : 0
  }
  function Pe(t, e) {
    return null != t.getExtension(e)
  }
  function Le(t) {
    try {
      if (null != jt(t)) return !0
    } catch (t) {
      return !1
    }
    return !1
  }
  function We(t) {
    if (0 === t) return !1
    var e = jt(t)
    if (1 === t) {
      if (!Pe(e, 'OES_texture_float')) return !1
    } else if (!Pe(e, 'EXT_color_buffer_float')) return !1
    return Ve(e)
  }
  function Ue(t) {
    if (0 === t) return !1
    var e = jt(t)
    if (1 !== t) {
      if (Pe(e, 'EXT_color_buffer_float')) return Ve(e)
      if (Pe(e, 'EXT_color_buffer_half_float')) {
        var n = e.getExtension('EXT_color_buffer_half_float')
        return (function (t, e) {
          var n = Qt(t, e),
            r = t.createTexture()
          t.bindTexture(t.TEXTURE_2D, r)
          t.texImage2D(
            t.TEXTURE_2D,
            0,
            n.internalFormatHalfFloat,
            1,
            1,
            0,
            n.textureFormatFloat,
            n.textureTypeHalfFloat,
            null
          )
          var o = t.createFramebuffer()
          t.bindFramebuffer(t.FRAMEBUFFER, o),
            t.framebufferTexture2D(
              t.FRAMEBUFFER,
              t.COLOR_ATTACHMENT0,
              t.TEXTURE_2D,
              r,
              0
            )
          var a =
            t.checkFramebufferStatus(t.FRAMEBUFFER) === t.FRAMEBUFFER_COMPLETE
          return (
            t.bindTexture(t.TEXTURE_2D, null),
            t.bindFramebuffer(t.FRAMEBUFFER, null),
            t.deleteTexture(r),
            t.deleteFramebuffer(o),
            a
          )
        })(e, n)
      }
      return !1
    }
    return (
      !!Pe(e, 'OES_texture_float') &&
      !!Pe(e, 'WEBGL_color_buffer_float') &&
      Ve(e)
    )
  }
  function Ve(t) {
    var e = Qt(t),
      n = t.createTexture()
    t.bindTexture(t.TEXTURE_2D, n)
    t.texImage2D(
      t.TEXTURE_2D,
      0,
      e.internalFormatFloat,
      1,
      1,
      0,
      e.textureFormatFloat,
      e.textureTypeFloat,
      null
    )
    var r = t.createFramebuffer()
    t.bindFramebuffer(t.FRAMEBUFFER, r),
      t.framebufferTexture2D(
        t.FRAMEBUFFER,
        t.COLOR_ATTACHMENT0,
        t.TEXTURE_2D,
        n,
        0
      )
    var o = t.checkFramebufferStatus(t.FRAMEBUFFER) === t.FRAMEBUFFER_COMPLETE
    return (
      t.bindTexture(t.TEXTURE_2D, null),
      t.bindFramebuffer(t.FRAMEBUFFER, null),
      t.deleteTexture(n),
      t.deleteFramebuffer(r),
      o
    )
  }
  function ze(t) {
    return 2 === t && null != jt(t).fenceSync
  }
  var Ge = Object.freeze({
      callAndCheck: Jt,
      canBeRepresented: ee,
      getWebGLErrorMessage: ne,
      getExtensionOrThrow: re,
      createVertexShader: oe,
      createFragmentShader: ae,
      createProgram: ce,
      linkProgram: le,
      validateProgram: he,
      createStaticVertexBuffer: fe,
      createStaticIndexBuffer: de,
      getNumChannels: function () {
        return 2 === i().getNumber('WEBGL_VERSION') ? 1 : 4
      },
      createTexture: pe,
      validateTextureSize: ve,
      createFramebuffer: me,
      bindVertexBufferToProgramAttribute: ge,
      bindTextureUnit: ye,
      unbindTextureUnit: function (t, e, n) {
        Se(t, n),
          Jt(t, e, function () {
            return t.activeTexture(t.TEXTURE0 + n)
          }),
          Jt(t, e, function () {
            return t.bindTexture(t.TEXTURE_2D, null)
          })
      },
      getProgramUniformLocationOrThrow: xe,
      getProgramUniformLocation: be,
      bindTextureToProgramUniformSampler: we,
      bindCanvasToFramebuffer: function (t, e) {
        Jt(t, e, function () {
          return t.bindFramebuffer(t.FRAMEBUFFER, null)
        }),
          Jt(t, e, function () {
            return t.viewport(0, 0, t.canvas.width, t.canvas.height)
          }),
          Jt(t, e, function () {
            return t.scissor(0, 0, t.canvas.width, t.canvas.height)
          })
      },
      bindColorTextureToFramebuffer: Ce,
      unbindColorTextureFromFramebuffer: Ee,
      validateFramebuffer: Re,
      getFramebufferErrorMessage: Ie,
      getBatchDim: Ae,
      getRowsCols: De,
      getShapeAs3D: Te,
      getTextureShapeFromLogicalShape: Ne,
      isReshapeFree: _e,
      getWebGLMaxTextureSize: Oe,
      resetMaxTextureSize: function () {
        ie = null
      },
      resetMaxTexturesInShader: function () {
        se = null
      },
      getMaxTexturesInShader: Me,
      getWebGLDisjointQueryTimerVersion: Be,
      hasExtension: Pe,
      isWebGLVersionEnabled: Le,
      isCapableOfRenderingToFloatTexture: We,
      isDownloadFloatTextureEnabled: Ue,
      isWebGLFenceEnabled: ze,
    }),
    He = i()
  function qe() {
    i().set('PROD', !0)
  }
  function Ke() {
    i().set('DEBUG', !0)
  }
  function je() {
    i().set('DEPRECATION_WARNINGS_ENABLED', !1),
      console.warn('TensorFlow.js deprecation warnings have been disabled.')
  }
  function Xe(t) {
    i().getBool('DEPRECATION_WARNINGS_ENABLED') &&
      console.warn(
        t +
          ' You can disable deprecation warnings with tf.disableDeprecationWarnings().'
      )
  }
  function Ye() {
    Lt.disposeVariables()
  }
  function $e() {
    return Lt
  }
  function Qe() {
    return Lt.memory()
  }
  function Je(t) {
    return Lt.profile(t)
  }
  function Ze(t, e) {
    return Lt.tidy(t, e)
  }
  function tn(t) {
    _t(t).forEach(function (t) {
      return t.dispose()
    })
  }
  function en(t) {
    return Lt.keep(t)
  }
  function nn(t) {
    return Lt.time(t)
  }
  function rn(t) {
    return Lt.setBackend(t)
  }
  function on() {
    return Lt.ready()
  }
  function an() {
    return Lt.backendName
  }
  function sn(t) {
    Lt.removeBackend(t)
  }
  function un(t) {
    return Lt.findBackend(t)
  }
  function cn(t) {
    return Lt.findBackendFactory(t)
  }
  function ln(t, e, n) {
    return void 0 === n && (n = 1), Lt.registerBackend(t, e, n)
  }
  function hn() {
    return Lt.backend
  }
  function fn(t, e) {
    i().setPlatform(t, e)
  }
  function dn() {
    for (var t = [], e = 0; e < arguments.length; e++) t[e] = arguments[e]
    i().getBool('IS_TEST') || console.warn.apply(console, t)
  }
  function pn(t, e) {
    var n = t
    if (V(t)) return 'string' === e ? [] : [t.length]
    if (!Array.isArray(t)) return []
    for (var r = []; Array.isArray(n) || (V(n) && 'string' !== e); )
      r.push(n.length), (n = n[0])
    return (
      Array.isArray(t) &&
        i().getBool('TENSORLIKE_CHECK_SHAPE_CONSISTENCY') &&
        (function t(e, n, r) {
          if (((r = r || []), !Array.isArray(e) && !V(e)))
            return void C(0 === n.length, function () {
              return (
                'Element arr[' +
                r.join('][') +
                '] is a primitive, but should be an array/TypedArray of ' +
                n[0] +
                ' elements'
              )
            })
          C(n.length > 0, function () {
            return (
              'Element arr[' +
              r.join('][') +
              '] should be a primitive, but is an array of ' +
              e.length +
              ' elements'
            )
          }),
            C(e.length === n[0], function () {
              return (
                'Element arr[' +
                r.join('][') +
                '] should have ' +
                n[0] +
                ' elements, but has ' +
                e.length +
                ' elements'
              )
            })
          for (var o = n.slice(1), a = 0; a < e.length; ++a)
            t(e[a], o, r.concat(a))
        })(t, r, []),
      r
    )
  }
  function vn(t, e, n, r) {
    if (
      null != t &&
      (('numeric' !== t && t !== e) || ('numeric' === t && 'string' === e))
    )
      throw new Error(
        "Argument '" +
          n +
          "' passed to '" +
          r +
          "' must be " +
          t +
          ' tensor, but got ' +
          e +
          ' tensor'
      )
  }
  function mn(t, e, n, r) {
    if ((void 0 === r && (r = 'numeric'), t instanceof wt))
      return vn(r, t.dtype, e, n), t
    var o = j(t)
    if (
      ('string' !== o &&
        ['bool', 'int32', 'float32'].indexOf(r) >= 0 &&
        (o = r),
      vn(r, o, e, n),
      null == t ||
        (!V(t) &&
          !Array.isArray(t) &&
          'number' != typeof t &&
          'boolean' != typeof t &&
          'string' != typeof t))
    ) {
      var a = null == t ? 'null' : t.constructor.name
      throw new Error(
        "Argument '" +
          e +
          "' passed to '" +
          n +
          "' must be a Tensor or TensorLike, but got '" +
          a +
          "'"
      )
    }
    var s = pn(t, o)
    V(t) || Array.isArray(t) || (t = [t])
    var u = 'string' !== o ? Q(t, o, i().getBool('DEBUG')) : I(t, [], !0)
    return Lt.makeTensor(u, s, o)
  }
  function gn(t, e, n, r) {
    if ((void 0 === r && (r = 'numeric'), !Array.isArray(t)))
      throw new Error(
        'Argument ' +
          e +
          ' passed to ' +
          n +
          ' must be a `Tensor[]` or `TensorLike[]`'
      )
    return t.map(function (t, r) {
      return mn(t, e + '[' + r + ']', n)
    }, r)
  }
  function yn(t, e) {
    for (var n = 0; n < t.length; ++n)
      if (t[t.length - n - 1] !== e - 1 - n) return !1
    return !0
  }
  function xn(t, e, n) {
    for (var r = t.length + e.length, o = [], a = 0, i = 0, s = 0; s < r; s++)
      -1 === n.indexOf(s) ? o.push(t[a++]) : o.push(e[i++])
    return o
  }
  function bn(t, e) {
    for (var n = [], r = t.length, o = 0; o < r; o++)
      -1 === e.indexOf(o) && n.push(t[o])
    return [
      n,
      e.map(function (e) {
        return t[e]
      }),
    ]
  }
  function wn(t, e) {
    return xn(
      t,
      e.map(function (t) {
        return 1
      }),
      e
    )
  }
  function Cn(t, e, n) {
    C(yn(e, n), function () {
      return (
        t +
        ' supports only inner-most axes for now. Got axes ' +
        e +
        ' and rank-' +
        n +
        ' input.'
      )
    })
  }
  function En(t, e) {
    if (yn(t, e)) return null
    for (var n = [], r = 0; r < e; ++r) -1 === t.indexOf(r) && n.push(r)
    return (
      t.forEach(function (t) {
        return n.push(t)
      }),
      n
    )
  }
  function Rn(t) {
    return t
      .map(function (t, e) {
        return [e, t]
      })
      .sort(function (t, e) {
        return t[1] - e[1]
      })
      .map(function (t) {
        return t[0]
      })
  }
  function In(t, e) {
    for (var n = [], r = e - t; r < e; ++r) n.push(r)
    return n
  }
  function kn(t, e) {
    var n = t[0].length
    t.forEach(function (t, e) {
      C(t.length === n, function () {
        return (
          'Error in concat' +
          n +
          'D: rank of tensors[' +
          e +
          '] must be the same as the rank of the rest (' +
          n +
          ')'
        )
      })
    }),
      C(e >= 0 && e < n, function () {
        return (
          'Error in concat' +
          n +
          'D: axis must be between 0 and ' +
          (n - 1) +
          '.'
        )
      })
    var r = t[0]
    t.forEach(function (t, o) {
      for (var a = 0; a < n; a++)
        C(a === e || t[a] === r[a], function () {
          return (
            'Error in concat' +
            n +
            'D: Shape of tensors[' +
            o +
            '] (' +
            t +
            ') does not match the shape of the rest (' +
            r +
            ') along the non-concatenated axis ' +
            o +
            '.'
          )
        })
    })
  }
  function Sn(t, e) {
    for (var n = t[0].slice(), r = 1; r < t.length; r++) n[e] += t[r][e]
    return n
  }
  function An(t) {
    var e = Object.keys(t)
    if (1 !== e.length)
      throw new Error(
        'Please provide an object with a single key (operation name) mapping to a function. Got an object with ' +
          e.length +
          ' keys.'
      )
    var n = e[0],
      r = t[n]
    n.endsWith('_') && (n = n.substring(0, n.length - 1))
    var o = function () {
      for (var t = [], e = 0; e < arguments.length; e++) t[e] = arguments[e]
      Lt.startScope(n)
      try {
        var o = r.apply(void 0, t)
        return (
          o instanceof Promise &&
            console.error('Cannot return a Promise inside of tidy.'),
          Lt.endScope(o),
          o
        )
      } catch (t) {
        throw (Lt.endScope(null), t)
      }
    }
    return Object.defineProperty(o, 'name', { value: n, configurable: !0 }), o
  }
  He.registerFlag('HAS_WEBGL', function () {
    return He.getNumber('WEBGL_VERSION') > 0
  }),
    He.registerFlag('WEBGL_VERSION', function () {
      return Le(2) ? 2 : Le(1) ? 1 : 0
    }),
    He.registerFlag('WEBGL_BUFFER_SUPPORTED', function () {
      return 2 === He.get('WEBGL_VERSION')
    }),
    He.registerFlag('WEBGL_CPU_FORWARD', function () {
      return !0
    }),
    He.registerFlag('WEBGL_FORCE_F16_TEXTURES', function () {
      return !1
    }),
    He.registerFlag('WEBGL_PACK', function () {
      return He.getBool('HAS_WEBGL')
    }),
    He.registerFlag('WEBGL_PACK_NORMALIZATION', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_PACK_CLIP', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_PACK_DEPTHWISECONV', function () {
      return !1
    }),
    He.registerFlag('WEBGL_PACK_BINARY_OPERATIONS', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_PACK_UNARY_OPERATIONS', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_PACK_ARRAY_OPERATIONS', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_PACK_IMAGE_OPERATIONS', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_PACK_REDUCE', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_LAZILY_UNPACK', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_CONV_IM2COL', function () {
      return He.getBool('WEBGL_PACK')
    }),
    He.registerFlag('WEBGL_MAX_TEXTURE_SIZE', function () {
      return Oe(He.getNumber('WEBGL_VERSION'))
    }),
    He.registerFlag('WEBGL_MAX_TEXTURES_IN_SHADER', function () {
      return Me(He.getNumber('WEBGL_VERSION'))
    }),
    He.registerFlag(
      'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION',
      function () {
        var t = He.getNumber('WEBGL_VERSION')
        return 0 === t ? 0 : Be(t)
      }
    ),
    He.registerFlag(
      'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE',
      function () {
        return (
          He.getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION') > 0 &&
          ((t = navigator.userAgent || navigator.vendor || window.opera),
          !(
            /(android|bb\d+|meego).+mobile|avantgo|bada\/|blackberry|blazer|compal|elaine|fennec|hiptop|iemobile|ip(hone|od)|iris|kindle|lge |maemo|midp|mmp|mobile.+firefox|netfront|opera m(ob|in)i|palm( os)?|phone|p(ixi|re)\/|plucker|pocket|psp|series(4|6)0|symbian|treo|up\.(browser|link)|vodafone|wap|windows ce|xda|xiino/i.test(
              t
            ) ||
            /1207|6310|6590|3gso|4thp|50[1-6]i|770s|802s|a wa|abac|ac(er|oo|s\-)|ai(ko|rn)|al(av|ca|co)|amoi|an(ex|ny|yw)|aptu|ar(ch|go)|as(te|us)|attw|au(di|\-m|r |s )|avan|be(ck|ll|nq)|bi(lb|rd)|bl(ac|az)|br(e|v)w|bumb|bw\-(n|u)|c55\/|capi|ccwa|cdm\-|cell|chtm|cldc|cmd\-|co(mp|nd)|craw|da(it|ll|ng)|dbte|dc\-s|devi|dica|dmob|do(c|p)o|ds(12|\-d)|el(49|ai)|em(l2|ul)|er(ic|k0)|esl8|ez([4-7]0|os|wa|ze)|fetc|fly(\-|_)|g1 u|g560|gene|gf\-5|g\-mo|go(\.w|od)|gr(ad|un)|haie|hcit|hd\-(m|p|t)|hei\-|hi(pt|ta)|hp( i|ip)|hs\-c|ht(c(\-| |_|a|g|p|s|t)|tp)|hu(aw|tc)|i\-(20|go|ma)|i230|iac( |\-|\/)|ibro|idea|ig01|ikom|im1k|inno|ipaq|iris|ja(t|v)a|jbro|jemu|jigs|kddi|keji|kgt( |\/)|klon|kpt |kwc\-|kyo(c|k)|le(no|xi)|lg( g|\/(k|l|u)|50|54|\-[a-w])|libw|lynx|m1\-w|m3ga|m50\/|ma(te|ui|xo)|mc(01|21|ca)|m\-cr|me(rc|ri)|mi(o8|oa|ts)|mmef|mo(01|02|bi|de|do|t(\-| |o|v)|zz)|mt(50|p1|v )|mwbp|mywa|n10[0-2]|n20[2-3]|n30(0|2)|n50(0|2|5)|n7(0(0|1)|10)|ne((c|m)\-|on|tf|wf|wg|wt)|nok(6|i)|nzph|o2im|op(ti|wv)|oran|owg1|p800|pan(a|d|t)|pdxg|pg(13|\-([1-8]|c))|phil|pire|pl(ay|uc)|pn\-2|po(ck|rt|se)|prox|psio|pt\-g|qa\-a|qc(07|12|21|32|60|\-[2-7]|i\-)|qtek|r380|r600|raks|rim9|ro(ve|zo)|s55\/|sa(ge|ma|mm|ms|ny|va)|sc(01|h\-|oo|p\-)|sdk\/|se(c(\-|0|1)|47|mc|nd|ri)|sgh\-|shar|sie(\-|m)|sk\-0|sl(45|id)|sm(al|ar|b3|it|t5)|so(ft|ny)|sp(01|h\-|v\-|v )|sy(01|mb)|t2(18|50)|t6(00|10|18)|ta(gt|lk)|tcl\-|tdg\-|tel(i|m)|tim\-|t\-mo|to(pl|sh)|ts(70|m\-|m3|m5)|tx\-9|up(\.b|g1|si)|utst|v400|v750|veri|vi(rg|te)|vk(40|5[0-3]|\-v)|vm40|voda|vulc|vx(52|53|60|61|70|80|81|83|85|98)|w3c(\-| )|webc|whit|wi(g |nc|nw)|wmlb|wonu|x700|yas\-|your|zeto|zte\-/i.test(
              t.substr(0, 4)
            )
          ))
        )
        var t
      }
    ),
    He.registerFlag('WEBGL_RENDER_FLOAT32_CAPABLE', function () {
      return We(He.getNumber('WEBGL_VERSION'))
    }),
    He.registerFlag('WEBGL_RENDER_FLOAT32_ENABLED', function () {
      return (
        !He.getBool('WEBGL_FORCE_F16_TEXTURES') &&
        He.getBool('WEBGL_RENDER_FLOAT32_CAPABLE')
      )
    }),
    He.registerFlag('WEBGL_DOWNLOAD_FLOAT_ENABLED', function () {
      return Ue(He.getNumber('WEBGL_VERSION'))
    }),
    He.registerFlag('WEBGL_FENCE_API_ENABLED', function () {
      return ze(He.getNumber('WEBGL_VERSION'))
    }),
    He.registerFlag('WEBGL_SIZE_UPLOAD_UNIFORM', function () {
      return He.getBool('WEBGL_RENDER_FLOAT32_ENABLED') ? 4 : 0
    }),
    (bt = Xe)
  var Dn = An({
      complex_: function (t, e) {
        var n = mn(t, 'real', 'complex'),
          r = mn(e, 'imag', 'complex')
        return (
          E(
            n.shape,
            r.shape,
            'real and imag shapes, ' +
              n.shape +
              ' and ' +
              r.shape +
              ', must match in call to tf.complex().'
          ),
          Lt.runKernelFunc(
            function (t) {
              return t.complex(n, r)
            },
            { $real: n, $imag: r }
          )
        )
      },
    }),
    Tn = An({
      real_: function (t) {
        var e = mn(t, 'input', 'real')
        return Lt.runKernelFunc(
          function (t) {
            return t.real(e)
          },
          { $input: e }
        )
      },
    }),
    Nn = An({
      imag_: function (t) {
        var e = mn(t, 'input', 'imag')
        return Lt.runKernelFunc(
          function (t) {
            return t.imag(e)
          },
          { $input: e }
        )
      },
    })
  function Fn(t, e, n) {
    return _n(t, e, pn(t, n), n)
  }
  function _n(t, e, n, r) {
    if ((null == r && (r = j(t)), 'complex64' === r))
      throw new Error(
        'Cannot construct a complex64 tensor directly. Please use tf.complex(real, imag).'
      )
    if (
      !V(t) &&
      !Array.isArray(t) &&
      'number' != typeof t &&
      'boolean' != typeof t &&
      'string' != typeof t
    )
      throw new Error(
        'values passed to tensor(values) must be a number/boolean/string or an array of numbers/booleans/strings, or a TypedArray'
      )
    if (null != e) {
      nt(e)
      var o = k(e),
        a = k(n)
      C(o === a, function () {
        return (
          'Based on the provided shape, [' +
          e +
          '], the tensor should have ' +
          o +
          ' values but has ' +
          a
        )
      })
      for (var s = 0; s < n.length; ++s) {
        var u = n[s],
          c = s !== n.length - 1 || u !== k(e.slice(s))
        C(n[s] === e[s] || !c, function () {
          return (
            'Error creating a new Tensor. Inferred shape (' +
            n +
            ') does not match the provided shape (' +
            e +
            '). '
          )
        })
      }
    }
    return (
      V(t) || Array.isArray(t) || (t = [t]),
      (e = e || n),
      (t = 'string' !== r ? Q(t, r, i().getBool('DEBUG')) : I(t, [], !0)),
      Lt.makeTensor(t, e, r)
    )
  }
  function On(t, e) {
    if (((V(t) && 'string' !== e) || Array.isArray(t)) && 'complex64' !== e)
      throw new Error(
        'Error creating a new Scalar: value must be a primitive (number|boolean|string)'
      )
    if ('string' === e && V(t) && !(t instanceof Uint8Array))
      throw new Error(
        'When making a scalar from encoded string, the value must be `Uint8Array`.'
      )
    return _n(t, [], [], e)
  }
  function Mn(t, e) {
    R(t)
    var n = pn(t, e)
    if (1 !== n.length)
      throw new Error('tensor1d() requires values to be a flat/TypedArray')
    return _n(t, null, n, e)
  }
  function Bn(t, e, n) {
    if ((R(t), null != e && 2 !== e.length))
      throw new Error('tensor2d() requires shape to have two numbers')
    var r = pn(t, n)
    if (2 !== r.length && 1 !== r.length)
      throw new Error(
        'tensor2d() requires values to be number[][] or flat/TypedArray'
      )
    if (1 === r.length && null == e)
      throw new Error(
        'tensor2d() requires shape to be provided when `values` are a flat/TypedArray'
      )
    return _n(t, e, r, n)
  }
  function Pn(t, e, n) {
    if ((R(t), null != e && 3 !== e.length))
      throw new Error('tensor3d() requires shape to have three numbers')
    var r = pn(t, n)
    if (3 !== r.length && 1 !== r.length)
      throw new Error(
        'tensor3d() requires values to be number[][][] or flat/TypedArray'
      )
    if (1 === r.length && null == e)
      throw new Error(
        'tensor3d() requires shape to be provided when `values` are a flat array'
      )
    return _n(t, e, r, n)
  }
  function Ln(t, e, n) {
    if ((R(t), null != e && 4 !== e.length))
      throw new Error('tensor4d() requires shape to have four numbers')
    var r = pn(t, n)
    if (4 !== r.length && 1 !== r.length)
      throw new Error(
        'tensor4d() requires values to be number[][][][] or flat/TypedArray'
      )
    if (1 === r.length && null == e)
      throw new Error(
        'tensor4d() requires shape to be provided when `values` are a flat array'
      )
    return _n(t, e, r, n)
  }
  function Wn(t, e, n) {
    if ((R(t), null != e && 5 !== e.length))
      throw new Error('tensor5d() requires shape to have five numbers')
    var r = pn(t, n)
    if (5 !== r.length && 1 !== r.length)
      throw new Error(
        'tensor5d() requires values to be number[][][][][] or flat/TypedArray'
      )
    if (1 === r.length && null == e)
      throw new Error(
        'tensor5d() requires shape to be provided when `values` are a flat array'
      )
    return _n(t, e, r, n)
  }
  function Un(t, e, n) {
    if ((R(t), null != e && 6 !== e.length))
      throw new Error('tensor6d() requires shape to have six numbers')
    var r = pn(t, n)
    if (6 !== r.length && 1 !== r.length)
      throw new Error(
        'tensor6d() requires values to be number[][][][][][] or flat/TypedArray'
      )
    if (1 === r.length && null == e)
      throw new Error(
        'tensor6d() requires shape to be provided when `values` are a flat array'
      )
    return _n(t, (e = e || r), r, n)
  }
  function Vn(t, e, n, r) {
    return void 0 === e && (e = !0), Lt.makeVariable(t, e, n, r)
  }
  function zn(t, e) {
    if ((void 0 === e && (e = 'float32'), 'complex64' === e)) {
      var n = zn(t, 'float32'),
        r = Gn(t, 'float32')
      return Dn(n, r)
    }
    var o = Z(k(t), e)
    return Lt.makeTensor(o, t, e)
  }
  function Gn(t, e) {
    if ((void 0 === e && (e = 'float32'), 'complex64' === e)) {
      var n = Gn(t, 'float32'),
        r = Gn(t, 'float32')
      return Dn(n, r)
    }
    var o = tt(k(t), e)
    return Lt.makeTensor(o, t, e)
  }
  function Hn(t, e, n) {
    return Lt.runKernelFunc(function (r) {
      return r.fill(t, e, n)
    }, {})
  }
  function qn(t, e, n) {
    if (n <= 0) throw new Error('The number of values should be positive.')
    return Lt.runKernelFunc(function (r) {
      return r.linspace(t, e, n)
    }, {})
  }
  function Kn(t, e, n, r) {
    if ((void 0 === n && (n = 1), void 0 === r && (r = 'float32'), 0 === n))
      throw new Error('Cannot have a step of zero')
    if (t === e || (t < e && n < 0) || (e < t && n > 1)) return Gn([0], r)
    var o = tt(Math.abs(Math.ceil((e - t) / n)), r)
    e < t && 1 === n && (n = -1), (o[0] = t)
    for (var a = 1; a < o.length; a++) o[a] = o[a - 1] + n
    return Mn(o, r)
  }
  var jn = An({
      onesLike_: function (t) {
        var e = mn(t, 'x', 'onesLike')
        if ('complex64' === e.dtype) {
          var n = jn(Tn(e)),
            r = Xn(Nn(e))
          return Dn(n, r)
        }
        return Lt.runKernelFunc(
          function (t) {
            return t.onesLike(e)
          },
          { $x: e },
          function (t, e) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Xn = An({
      zerosLike_: function (t) {
        var e = mn(t, 'x', 'zerosLike')
        return Lt.runKernelFunc(
          function (t) {
            return t.zerosLike(e)
          },
          { $x: e },
          function (t, e) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    })
  var Yn = An({
      concat_: function (t, e) {
        void 0 === e && (e = 0),
          C(t.length >= 1, function () {
            return 'Pass at least one tensor to concat'
          })
        var n = gn(t, 'tensors', 'concat')
        'complex64' === n[0].dtype &&
          n.forEach(function (t) {
            if ('complex64' !== t.dtype)
              throw new Error(
                'Cannot concatenate complex64 tensors with a tensor\n          with dtype ' +
                  t.dtype +
                  '. '
              )
          }),
          (e = O(e, n[0].shape)[0])
        var r = Sn(
          n.map(function (t) {
            return t.shape
          }),
          e
        )
        if (0 === k(r)) return Fn([], r)
        if (
          1 ===
          (n = n.filter(function (t) {
            return t.size > 0
          })).length
        )
          return n[0]
        var o = n.map(function (t) {
          return t.shape
        })
        kn(o, e)
        var a = n,
          i = { axis: e }
        return Lt.runKernelFunc(
          function (t) {
            return t.concat(n, e)
          },
          a,
          function (t) {
            var n = o.map(function (t) {
              return t[e]
            })
            return tr(t, n, e).map(function (t) {
              return function () {
                return t
              }
            })
          },
          'Concat',
          i
        )
      },
    }),
    $n = An({
      concat1d_: function (t) {
        return Yn(t, 0)
      },
    }),
    Qn = An({
      concat2d_: function (t, e) {
        return Yn(t, e)
      },
    }),
    Jn = An({
      concat3d_: function (t, e) {
        return Yn(t, e)
      },
    }),
    Zn = An({
      concat4d_: function (t, e) {
        return Yn(t, e)
      },
    }),
    tr = An({
      split_: function (t, e, n) {
        void 0 === n && (n = 0)
        var r,
          o = mn(t, 'x', 'split')
        return (
          (n = O(n, o.shape)[0]),
          'number' == typeof e
            ? (C(o.shape[n] % e == 0, function () {
                return 'Number of splits must evenly divide the axis.'
              }),
              (r = new Array(e).fill(o.shape[n] / e)))
            : (C(
                o.shape[n] ===
                  e.reduce(function (t, e) {
                    return t + e
                  }),
                function () {
                  return 'The sum of sizes must match the size of the axis dimension.'
                }
              ),
              (r = e)),
          Lt.runKernelFunc(
            function (t) {
              return t.split(o, r, n)
            },
            { $x: o },
            function (t) {
              return {
                $x: function () {
                  return Yn(t, n)
                },
              }
            }
          )
        )
      },
    })
  function er(t, e) {
    return t((e = { exports: {} }), e.exports), e.exports
  }
  var nr = er(function (t) {
      !(function (t, e, n) {
        function r(t) {
          var e,
            n = this,
            r =
              ((e = 4022871197),
              function (t) {
                t = t.toString()
                for (var n = 0; n < t.length; n++) {
                  var r = 0.02519603282416938 * (e += t.charCodeAt(n))
                  ;(r -= e = r >>> 0),
                    (e = (r *= e) >>> 0),
                    (e += 4294967296 * (r -= e))
                }
                return 2.3283064365386963e-10 * (e >>> 0)
              })
          ;(n.next = function () {
            var t = 2091639 * n.s0 + 2.3283064365386963e-10 * n.c
            return (n.s0 = n.s1), (n.s1 = n.s2), (n.s2 = t - (n.c = 0 | t))
          }),
            (n.c = 1),
            (n.s0 = r(' ')),
            (n.s1 = r(' ')),
            (n.s2 = r(' ')),
            (n.s0 -= r(t)),
            n.s0 < 0 && (n.s0 += 1),
            (n.s1 -= r(t)),
            n.s1 < 0 && (n.s1 += 1),
            (n.s2 -= r(t)),
            n.s2 < 0 && (n.s2 += 1),
            (r = null)
        }
        function o(t, e) {
          return (e.c = t.c), (e.s0 = t.s0), (e.s1 = t.s1), (e.s2 = t.s2), e
        }
        function a(t, e) {
          var n = new r(t),
            a = e && e.state,
            i = n.next
          return (
            (i.int32 = function () {
              return (4294967296 * n.next()) | 0
            }),
            (i.double = function () {
              return i() + 11102230246251565e-32 * ((2097152 * i()) | 0)
            }),
            (i.quick = i),
            a &&
              ('object' == typeof a && o(a, n),
              (i.state = function () {
                return o(n, {})
              })),
            i
          )
        }
        e && e.exports
          ? (e.exports = a)
          : n && n.amd
          ? n(function () {
              return a
            })
          : (this.alea = a)
      })(0, t, !1)
    }),
    rr = er(function (t) {
      !(function (t, e, n) {
        function r(t) {
          var e = this,
            n = ''
          ;(e.x = 0),
            (e.y = 0),
            (e.z = 0),
            (e.w = 0),
            (e.next = function () {
              var t = e.x ^ (e.x << 11)
              return (
                (e.x = e.y),
                (e.y = e.z),
                (e.z = e.w),
                (e.w ^= (e.w >>> 19) ^ t ^ (t >>> 8))
              )
            }),
            t === (0 | t) ? (e.x = t) : (n += t)
          for (var r = 0; r < n.length + 64; r++)
            (e.x ^= 0 | n.charCodeAt(r)), e.next()
        }
        function o(t, e) {
          return (e.x = t.x), (e.y = t.y), (e.z = t.z), (e.w = t.w), e
        }
        function a(t, e) {
          var n = new r(t),
            a = e && e.state,
            i = function () {
              return (n.next() >>> 0) / 4294967296
            }
          return (
            (i.double = function () {
              do {
                var t =
                  ((n.next() >>> 11) + (n.next() >>> 0) / 4294967296) /
                  (1 << 21)
              } while (0 === t)
              return t
            }),
            (i.int32 = n.next),
            (i.quick = i),
            a &&
              ('object' == typeof a && o(a, n),
              (i.state = function () {
                return o(n, {})
              })),
            i
          )
        }
        e && e.exports
          ? (e.exports = a)
          : n && n.amd
          ? n(function () {
              return a
            })
          : (this.xor128 = a)
      })(0, t, !1)
    }),
    or = er(function (t) {
      !(function (t, e, n) {
        function r(t) {
          var e = this,
            n = ''
          ;(e.next = function () {
            var t = e.x ^ (e.x >>> 2)
            return (
              (e.x = e.y),
              (e.y = e.z),
              (e.z = e.w),
              (e.w = e.v),
              ((e.d = (e.d + 362437) | 0) +
                (e.v = e.v ^ (e.v << 4) ^ t ^ (t << 1))) |
                0
            )
          }),
            (e.x = 0),
            (e.y = 0),
            (e.z = 0),
            (e.w = 0),
            (e.v = 0),
            t === (0 | t) ? (e.x = t) : (n += t)
          for (var r = 0; r < n.length + 64; r++)
            (e.x ^= 0 | n.charCodeAt(r)),
              r == n.length && (e.d = (e.x << 10) ^ (e.x >>> 4)),
              e.next()
        }
        function o(t, e) {
          return (
            (e.x = t.x),
            (e.y = t.y),
            (e.z = t.z),
            (e.w = t.w),
            (e.v = t.v),
            (e.d = t.d),
            e
          )
        }
        function a(t, e) {
          var n = new r(t),
            a = e && e.state,
            i = function () {
              return (n.next() >>> 0) / 4294967296
            }
          return (
            (i.double = function () {
              do {
                var t =
                  ((n.next() >>> 11) + (n.next() >>> 0) / 4294967296) /
                  (1 << 21)
              } while (0 === t)
              return t
            }),
            (i.int32 = n.next),
            (i.quick = i),
            a &&
              ('object' == typeof a && o(a, n),
              (i.state = function () {
                return o(n, {})
              })),
            i
          )
        }
        e && e.exports
          ? (e.exports = a)
          : n && n.amd
          ? n(function () {
              return a
            })
          : (this.xorwow = a)
      })(0, t, !1)
    }),
    ar = er(function (t) {
      !(function (t, e, n) {
        function r(t) {
          var e = this
          ;(e.next = function () {
            var t,
              n,
              r = e.x,
              o = e.i
            return (
              (t = r[o]),
              (n = (t ^= t >>> 7) ^ (t << 24)),
              (n ^= (t = r[(o + 1) & 7]) ^ (t >>> 10)),
              (n ^= (t = r[(o + 3) & 7]) ^ (t >>> 3)),
              (n ^= (t = r[(o + 4) & 7]) ^ (t << 7)),
              (t = r[(o + 7) & 7]),
              (n ^= (t ^= t << 13) ^ (t << 9)),
              (r[o] = n),
              (e.i = (o + 1) & 7),
              n
            )
          }),
            (function (t, e) {
              var n,
                r = []
              if (e === (0 | e)) r[0] = e
              else
                for (e = '' + e, n = 0; n < e.length; ++n)
                  r[7 & n] =
                    (r[7 & n] << 15) ^
                    ((e.charCodeAt(n) + r[(n + 1) & 7]) << 13)
              for (; r.length < 8; ) r.push(0)
              for (n = 0; n < 8 && 0 === r[n]; ++n);
              for (
                8 == n ? (r[7] = -1) : r[n], t.x = r, t.i = 0, n = 256;
                n > 0;
                --n
              )
                t.next()
            })(e, t)
        }
        function o(t, e) {
          return (e.x = t.x.slice()), (e.i = t.i), e
        }
        function a(t, e) {
          null == t && (t = +new Date())
          var n = new r(t),
            a = e && e.state,
            i = function () {
              return (n.next() >>> 0) / 4294967296
            }
          return (
            (i.double = function () {
              do {
                var t =
                  ((n.next() >>> 11) + (n.next() >>> 0) / 4294967296) /
                  (1 << 21)
              } while (0 === t)
              return t
            }),
            (i.int32 = n.next),
            (i.quick = i),
            a &&
              (a.x && o(a, n),
              (i.state = function () {
                return o(n, {})
              })),
            i
          )
        }
        e && e.exports
          ? (e.exports = a)
          : n && n.amd
          ? n(function () {
              return a
            })
          : (this.xorshift7 = a)
      })(0, t, !1)
    }),
    ir = er(function (t) {
      !(function (t, e, n) {
        function r(t) {
          var e = this
          ;(e.next = function () {
            var t,
              n,
              r = e.w,
              o = e.X,
              a = e.i
            return (
              (e.w = r = (r + 1640531527) | 0),
              (n = o[(a + 34) & 127]),
              (t = o[(a = (a + 1) & 127)]),
              (n ^= n << 13),
              (t ^= t << 17),
              (n ^= n >>> 15),
              (t ^= t >>> 12),
              (n = o[a] = n ^ t),
              (e.i = a),
              (n + (r ^ (r >>> 16))) | 0
            )
          }),
            (function (t, e) {
              var n,
                r,
                o,
                a,
                i,
                s = [],
                u = 128
              for (
                e === (0 | e)
                  ? ((r = e), (e = null))
                  : ((e += '\0'), (r = 0), (u = Math.max(u, e.length))),
                  o = 0,
                  a = -32;
                a < u;
                ++a
              )
                e && (r ^= e.charCodeAt((a + 32) % e.length)),
                  0 === a && (i = r),
                  (r ^= r << 10),
                  (r ^= r >>> 15),
                  (r ^= r << 4),
                  (r ^= r >>> 13),
                  a >= 0 &&
                    ((i = (i + 1640531527) | 0),
                    (o = 0 == (n = s[127 & a] ^= r + i) ? o + 1 : 0))
              for (
                o >= 128 && (s[127 & ((e && e.length) || 0)] = -1),
                  o = 127,
                  a = 512;
                a > 0;
                --a
              )
                (r = s[(o + 34) & 127]),
                  (n = s[(o = (o + 1) & 127)]),
                  (r ^= r << 13),
                  (n ^= n << 17),
                  (r ^= r >>> 15),
                  (n ^= n >>> 12),
                  (s[o] = r ^ n)
              ;(t.w = i), (t.X = s), (t.i = o)
            })(e, t)
        }
        function o(t, e) {
          return (e.i = t.i), (e.w = t.w), (e.X = t.X.slice()), e
        }
        function a(t, e) {
          null == t && (t = +new Date())
          var n = new r(t),
            a = e && e.state,
            i = function () {
              return (n.next() >>> 0) / 4294967296
            }
          return (
            (i.double = function () {
              do {
                var t =
                  ((n.next() >>> 11) + (n.next() >>> 0) / 4294967296) /
                  (1 << 21)
              } while (0 === t)
              return t
            }),
            (i.int32 = n.next),
            (i.quick = i),
            a &&
              (a.X && o(a, n),
              (i.state = function () {
                return o(n, {})
              })),
            i
          )
        }
        e && e.exports
          ? (e.exports = a)
          : n && n.amd
          ? n(function () {
              return a
            })
          : (this.xor4096 = a)
      })(0, t, !1)
    }),
    sr = er(function (t) {
      !(function (t, e, n) {
        function r(t) {
          var e = this,
            n = ''
          ;(e.next = function () {
            var t = e.b,
              n = e.c,
              r = e.d,
              o = e.a
            return (
              (t = (t << 25) ^ (t >>> 7) ^ n),
              (n = (n - r) | 0),
              (r = (r << 24) ^ (r >>> 8) ^ o),
              (o = (o - t) | 0),
              (e.b = t = (t << 20) ^ (t >>> 12) ^ n),
              (e.c = n = (n - r) | 0),
              (e.d = (r << 16) ^ (n >>> 16) ^ o),
              (e.a = (o - t) | 0)
            )
          }),
            (e.a = 0),
            (e.b = 0),
            (e.c = -1640531527),
            (e.d = 1367130551),
            t === Math.floor(t)
              ? ((e.a = (t / 4294967296) | 0), (e.b = 0 | t))
              : (n += t)
          for (var r = 0; r < n.length + 20; r++)
            (e.b ^= 0 | n.charCodeAt(r)), e.next()
        }
        function o(t, e) {
          return (e.a = t.a), (e.b = t.b), (e.c = t.c), (e.d = t.d), e
        }
        function a(t, e) {
          var n = new r(t),
            a = e && e.state,
            i = function () {
              return (n.next() >>> 0) / 4294967296
            }
          return (
            (i.double = function () {
              do {
                var t =
                  ((n.next() >>> 11) + (n.next() >>> 0) / 4294967296) /
                  (1 << 21)
              } while (0 === t)
              return t
            }),
            (i.int32 = n.next),
            (i.quick = i),
            a &&
              ('object' == typeof a && o(a, n),
              (i.state = function () {
                return o(n, {})
              })),
            i
          )
        }
        e && e.exports
          ? (e.exports = a)
          : n && n.amd
          ? n(function () {
              return a
            })
          : (this.tychei = a)
      })(0, t, !1)
    }),
    ur = er(function (t) {
      !(function (e, n) {
        var r,
          o = this,
          a = 256,
          i = 6,
          s = 'random',
          u = n.pow(a, i),
          c = n.pow(2, 52),
          l = 2 * c,
          h = a - 1
        function f(t, h, f) {
          var g = [],
            y = v(
              (function t(e, n) {
                var r,
                  o = [],
                  a = typeof e
                if (n && 'object' == a)
                  for (r in e)
                    try {
                      o.push(t(e[r], n - 1))
                    } catch (t) {}
                return o.length ? o : 'string' == a ? e : e + '\0'
              })(
                (h = 1 == h ? { entropy: !0 } : h || {}).entropy
                  ? [t, m(e)]
                  : null == t
                  ? (function () {
                      try {
                        var t
                        return (
                          r && (t = r.randomBytes)
                            ? (t = t(a))
                            : ((t = new Uint8Array(a)),
                              (o.crypto || o.msCrypto).getRandomValues(t)),
                          m(t)
                        )
                      } catch (t) {
                        var n = o.navigator,
                          i = n && n.plugins
                        return [+new Date(), o, i, o.screen, m(e)]
                      }
                    })()
                  : t,
                3
              ),
              g
            ),
            x = new d(g),
            b = function () {
              for (var t = x.g(i), e = u, n = 0; t < c; )
                (t = (t + n) * a), (e *= a), (n = x.g(1))
              for (; t >= l; ) (t /= 2), (e /= 2), (n >>>= 1)
              return (t + n) / e
            }
          return (
            (b.int32 = function () {
              return 0 | x.g(4)
            }),
            (b.quick = function () {
              return x.g(4) / 4294967296
            }),
            (b.double = b),
            v(m(x.S), e),
            (
              h.pass ||
              f ||
              function (t, e, r, o) {
                return (
                  o &&
                    (o.S && p(o, x),
                    (t.state = function () {
                      return p(x, {})
                    })),
                  r ? ((n[s] = t), e) : t
                )
              }
            )(b, y, 'global' in h ? h.global : this == n, h.state)
          )
        }
        function d(t) {
          var e,
            n = t.length,
            r = this,
            o = 0,
            i = (r.i = r.j = 0),
            s = (r.S = [])
          for (n || (t = [n++]); o < a; ) s[o] = o++
          for (o = 0; o < a; o++)
            (s[o] = s[(i = h & (i + t[o % n] + (e = s[o])))]), (s[i] = e)
          ;(r.g = function (t) {
            for (var e, n = 0, o = r.i, i = r.j, s = r.S; t--; )
              (e = s[(o = h & (o + 1))]),
                (n =
                  n * a + s[h & ((s[o] = s[(i = h & (i + e))]) + (s[i] = e))])
            return (r.i = o), (r.j = i), n
          })(a)
        }
        function p(t, e) {
          return (e.i = t.i), (e.j = t.j), (e.S = t.S.slice()), e
        }
        function v(t, e) {
          for (var n, r = t + '', o = 0; o < r.length; )
            e[h & o] = h & ((n ^= 19 * e[h & o]) + r.charCodeAt(o++))
          return m(e)
        }
        function m(t) {
          return String.fromCharCode.apply(0, t)
        }
        if (((n['seed' + s] = f), v(n.random(), e), t.exports)) {
          t.exports = f
          try {
            r = require('crypto')
          } catch (t) {}
        }
      })([], Math)
    })
  ;(ur.alea = nr),
    (ur.xor128 = rr),
    (ur.xorwow = or),
    (ur.xorshift7 = ar),
    (ur.xor4096 = ir),
    (ur.tychei = sr)
  var cr = ur.alea,
    lr = (function () {
      function t(t, e, n, r, o) {
        ;(this.mean = t),
          (this.stdDev = e),
          (this.dtype = n),
          (this.nextVal = NaN),
          (this.truncated = r),
          this.truncated &&
            ((this.upper = this.mean + 2 * this.stdDev),
            (this.lower = this.mean - 2 * this.stdDev))
        var a = o || Math.random()
        this.random = cr(a.toString())
      }
      return (
        (t.prototype.nextValue = function () {
          if (!isNaN(this.nextVal)) {
            var t = this.nextVal
            return (this.nextVal = NaN), t
          }
          for (var e, n, r = !1; !r; ) {
            var o = void 0,
              a = void 0,
              i = void 0
            do {
              i =
                (o = 2 * this.random() - 1) * o +
                (a = 2 * this.random() - 1) * a
            } while (i >= 1 || 0 === i)
            var s = Math.sqrt((-2 * Math.log(i)) / i)
            ;(e = this.mean + this.stdDev * o * s),
              (n = this.mean + this.stdDev * a * s),
              (this.truncated && !this.isValidTruncated(e)) || (r = !0)
          }
          return (
            (this.truncated && !this.isValidTruncated(n)) ||
              (this.nextVal = this.convertValue(n)),
            this.convertValue(e)
          )
        }),
        (t.prototype.convertValue = function (t) {
          return null == this.dtype || 'float32' === this.dtype
            ? t
            : Math.round(t)
        }),
        (t.prototype.isValidTruncated = function (t) {
          return t <= this.upper && t >= this.lower
        }),
        t
      )
    })(),
    hr = (function () {
      function t(t, e, n, r) {
        ;(this.alpha = t), (this.beta = 1 / e), (this.dtype = n)
        var o = r || Math.random()
        ;(this.randu = cr(o.toString())),
          (this.randn = new lr(0, 1, n, !1, this.randu())),
          (this.d = t < 1 ? t + 2 / 3 : t - 1 / 3),
          (this.c = 1 / Math.sqrt(9 * this.d))
      }
      return (
        (t.prototype.nextValue = function () {
          for (var t, e, n, r, o, a; ; ) {
            do {
              ;(r = this.randn.nextValue()), (a = 1 + this.c * r)
            } while (a <= 0)
            if (
              ((a *= a * a),
              (e = 1 - 0.331 * (t = r * r) * t),
              (n = 0.5 * t + this.d * (1 - a + Math.log(a))),
              (o = this.randu()) < e || Math.log(o) < n)
            )
              break
          }
          return (
            (a = (1 / this.beta) * this.d * a),
            this.alpha < 1 && (a *= Math.pow(this.randu(), 1 / this.alpha)),
            this.convertValue(a)
          )
        }),
        (t.prototype.convertValue = function (t) {
          return 'float32' === this.dtype ? t : Math.round(t)
        }),
        t
      )
    })(),
    fr = (function () {
      function t(t, e, n, r) {
        var o = this
        if (
          (void 0 === t && (t = 0),
          void 0 === e && (e = 1),
          (this.canReturnFloat = function () {
            return null == o.dtype || 'float32' === o.dtype
          }),
          (this.min = t),
          (this.range = e - t),
          (this.dtype = n),
          null == r && (r = Math.random()),
          'number' == typeof r && (r = r.toString()),
          !this.canReturnFloat() && this.range <= 1)
        )
          throw new Error(
            'The difference between ' +
              t +
              ' - ' +
              e +
              ' <= 1 and dtype is not float'
          )
        this.random = cr(r)
      }
      return (
        (t.prototype.convertValue = function (t) {
          return this.canReturnFloat() ? t : Math.round(t)
        }),
        (t.prototype.nextValue = function () {
          return this.convertValue(this.min + this.range * this.random())
        }),
        t
      )
    })()
  function dr(t, e, n) {
    return (
      void 0 === e && (e = 'float32'),
      (e = e || 'float32'),
      nt(t),
      new gt(t, e, n)
    )
  }
  function pr(t, e) {
    void 0 === e && (e = !1), console.log(t.toString(e))
  }
  var vr = An({
      batchToSpaceND_: function (t, e, n) {
        var r = mn(t, 'x', 'batchToSpaceND'),
          o = e.reduce(function (t, e) {
            return t * e
          })
        return (
          C(r.rank >= 1 + e.length, function () {
            return (
              'input rank is ' +
              r.rank +
              ' but should be > than blockShape.length ' +
              e.length
            )
          }),
          C(n.length === e.length, function () {
            return (
              'crops.length is ' +
              n.length +
              ' but should be equal to blockShape.length  ' +
              e.length
            )
          }),
          C(r.shape[0] % o == 0, function () {
            return (
              'input tensor batch is ' +
              r.shape[0] +
              ' but is not divisible by the product of the elements of blockShape ' +
              e.join(' * ') +
              ' === ' +
              o
            )
          }),
          Lt.runKernelFunc(
            function (t) {
              return t.batchToSpaceND(r, e, n)
            },
            { $x: r },
            function (t) {
              return {
                $x: function () {
                  return t.spaceToBatchND(e, n)
                },
              }
            }
          )
        )
      },
    }),
    mr = An({
      broadcastTo_: function (t, e) {
        var n = mn(t, 'broadcastTo', 'x'),
          r = n.shape
        if (
          e.some(function (t) {
            return !(t > 0) || t % 1 != 0
          })
        )
          throw new Error('broadcastTo(): Invalid broadcast shape [' + e + '].')
        if (e.length < n.rank)
          throw new Error(
            'broadcastTo(): shape.length=' +
              e.length +
              ' < input.rank=' +
              n.rank +
              '.'
          )
        if (e.length > n.rank) {
          for (var o = n.shape.slice(); o.length < e.length; ) o.unshift(1)
          n = n.reshape(o)
        }
        for (var a = Array.from(e), i = e.length - 1; i >= 0; i--)
          if (n.shape[i] === e[i]) a[i] = 1
          else if (1 !== n.shape[i])
            throw new Error(
              'broadcastTo(): [' + r + '] cannot be broadcast to [' + e + '].'
            )
        var s = a
          .map(function (t, e) {
            return t > 1 ? e : -1
          })
          .filter(function (t) {
            return t >= 0
          })
        return 0 === s.length
          ? n.clone()
          : Lt.runKernelFunc(
              function (t) {
                return t.tile(n, a)
              },
              { input: n },
              function (t) {
                return {
                  input: function () {
                    return t.sum(s, !0)
                  },
                }
              }
            )
      },
    }),
    gr = An({
      cast_: function (t, e) {
        var n = mn(t, 'x', 'cast')
        if (!W(e)) throw new Error('Failed to cast to unknown dtype ' + e)
        if (
          ('string' === e && 'string' !== n.dtype) ||
          ('string' !== e && 'string' === n.dtype)
        )
          throw new Error('Only strings can be casted to strings')
        var r = { dtype: e }
        return Lt.runKernelFunc(
          function (t) {
            return t.cast(n, e)
          },
          { x: n },
          function (t) {
            return {
              x: function () {
                return t.clone()
              },
            }
          },
          'Cast',
          r
        )
      },
    }),
    yr = An({
      clone_: function (t) {
        var e = mn(t, 'x', 'clone', null)
        return Lt.runKernelFunc(
          function () {
            return Lt.makeTensorFromDataId(e.dataId, e.shape, e.dtype)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return t.toFloat()
              },
            }
          }
        )
      },
    }),
    xr = An({
      cumsum_: function (t, e, n, r) {
        void 0 === e && (e = 0),
          void 0 === n && (n = !1),
          void 0 === r && (r = !1)
        var o = mn(t, 'x', 'cumsum'),
          a = En([(e |= 0)], o.rank),
          i = o
        null != a && (i = o.transpose(a))
        var s = In(1, o.rank)[0],
          u = Lt.runKernelFunc(
            function (t) {
              return t.cumsum(i, s, n, r)
            },
            { permutedX: i },
            function (t) {
              return {
                permutedX: function () {
                  return t.cumsum(e, n, !r)
                },
              }
            }
          )
        return null != a && (u = u.transpose(a)), u
      },
    }),
    br = An({
      depthToSpace_: function (t, e, n) {
        void 0 === n && (n = 'NHWC')
        var r = mn(t, 'x', 'depthToSpace'),
          o = 'NHWC' === n ? r.shape[1] : r.shape[2],
          a = 'NHWC' === n ? r.shape[2] : r.shape[3],
          i = 'NHWC' === n ? r.shape[3] : r.shape[1]
        return (
          C(o * e >= 0, function () {
            return (
              'Negative dimension size caused by overflow when multiplying\n      ' +
              o +
              ' and ' +
              e +
              '  for depthToSpace with input shape\n      ' +
              r.shape
            )
          }),
          C(a * e >= 0, function () {
            return (
              'Negative dimension size caused by overflow when multiplying\n      ' +
              a +
              ' and ' +
              e +
              ' for depthToSpace with input shape\n          ' +
              r.shape
            )
          }),
          C(i % (e * e) == 0, function () {
            return (
              'Dimension size must be evenly divisible by ' +
              e * e +
              ' but is ' +
              i +
              ' for depthToSpace with input shape ' +
              r.shape
            )
          }),
          Lt.runKernelFunc(
            function (t) {
              return t.depthToSpace(r, e, n)
            },
            { $x: r }
          )
        )
      },
    }),
    wr = An({
      expandDims_: function (t, e) {
        void 0 === e && (e = 0)
        var n = mn(t, 'x', 'expandDims', null)
        C(e <= n.rank, function () {
          return 'Axis must be <= rank of the tensor'
        })
        var r = n.shape.slice()
        return (
          e < 0 &&
            (C(-(n.rank + 1) <= e, function () {
              return (
                'Axis must be in the interval [' +
                -(n.rank + 1) +
                ', ' +
                n.rank +
                ']'
              )
            }),
            (e = n.rank + e + 1)),
          r.splice(e, 0, 1),
          Or(n, r)
        )
      },
    }),
    Cr = An({
      eye_: function (t, e, n, r) {
        void 0 === r && (r = 'float32'), null == e && (e = t)
        for (var o = dr([t, e], r), a = t <= e ? t : e, i = 0; i < a; ++i)
          o.set(1, i, i)
        var s = o.toTensor().as2D(t, e)
        if (null == n) return s
        if (1 === n.length) return Lr(wr(s, 0), [n[0], 1, 1])
        if (2 === n.length) return Lr(wr(wr(s, 0), 0), [n[0], n[1], 1, 1])
        if (3 === n.length)
          return Lr(wr(wr(wr(s, 0), 0), 0), [n[0], n[1], n[2], 1, 1])
        throw new Error(
          'eye() currently supports only 1D and 2D batchShapes, but received ' +
            n.length +
            'D.'
        )
      },
    }),
    Er = An({
      multinomial_: function (t, e, n, r) {
        void 0 === r && (r = !1)
        var o = mn(t, 'logits', 'multinomial'),
          a = o.size,
          i = o.rank
        if (a < 2)
          throw new Error(
            'Error in multinomial: you need at least 2 outcomes, but got ' +
              a +
              '.'
          )
        if (i > 2)
          throw new Error('Rank of probabilities must be 1 or 2, but is ' + i)
        n = n || Math.random()
        var s = 1 === i ? o.as2D(1, -1) : o,
          u = Lt.runKernelFunc(
            function (t) {
              return t.multinomial(s, r, e, n)
            },
            { logits2D: s }
          )
        return 1 === i ? u.as1D() : u
      },
    }),
    Rr = An({
      oneHot_: function (t, e, n, r) {
        if ((void 0 === n && (n = 1), void 0 === r && (r = 0), e < 2))
          throw new Error('Error in oneHot: depth must be >=2, but it is ' + e)
        var o = mn(t, 'indices', 'oneHot', 'int32'),
          a = o.shape.concat([e])
        return (
          (o = o.flatten()),
          Lt.runKernelFunc(
            function (t) {
              return t.oneHot(o, e, n, r)
            },
            { $indices: o },
            function (t) {
              return {
                $indices: function () {
                  return Gn(o.shape, 'float32')
                },
              }
            }
          ).reshape(a)
        )
      },
    }),
    Ir = An({
      pad_: function (t, e, n) {
        void 0 === n && (n = 0)
        var r = mn(t, 'x', 'pad')
        if (0 === r.rank)
          throw new Error('pad(scalar) is not defined. Pass non-scalar to pad')
        var o = { paddings: e, constantValue: n }
        return Lt.runKernelFunc(
          function (t) {
            return t.pad(r, e, n)
          },
          { x: r },
          function (t) {
            var n = e.map(function (t) {
              return t[0]
            })
            return {
              x: function () {
                return t.slice(n, r.shape)
              },
            }
          },
          'PadV2',
          o
        )
      },
    }),
    kr = An({
      pad1d_: function (t, e, n) {
        return (
          void 0 === n && (n = 0),
          C(2 === e.length, function () {
            return 'Invalid number of paddings. Must be length of 2.'
          }),
          Ir(t, [e], n)
        )
      },
    }),
    Sr = An({
      pad2d_: function (t, e, n) {
        return (
          void 0 === n && (n = 0),
          C(
            2 === e.length && 2 === e[0].length && 2 === e[1].length,
            function () {
              return 'Invalid number of paddings. Must be length of 2 each.'
            }
          ),
          Ir(t, e, n)
        )
      },
    }),
    Ar = An({
      pad3d_: function (t, e, n) {
        return (
          void 0 === n && (n = 0),
          C(
            3 === e.length &&
              2 === e[0].length &&
              2 === e[1].length &&
              2 === e[2].length,
            function () {
              return 'Invalid number of paddings. Must be length of 2 each.'
            }
          ),
          Ir(t, e, n)
        )
      },
    }),
    Dr = An({
      pad4d_: function (t, e, n) {
        return (
          void 0 === n && (n = 0),
          C(
            4 === e.length &&
              2 === e[0].length &&
              2 === e[1].length &&
              2 === e[2].length &&
              2 === e[3].length,
            function () {
              return 'Invalid number of paddings. Must be length of 2 each.'
            }
          ),
          Ir(t, e, n)
        )
      },
    }),
    Tr = An({
      rand_: function (t, e, n) {
        var r = k(t),
          o = null
        if (null == n || 'float32' === n) o = new Float32Array(r)
        else if ('int32' === n) o = new Int32Array(r)
        else {
          if ('bool' !== n) throw new Error('Unknown data type ' + n)
          o = new Uint8Array(r)
        }
        for (var a = 0; a < r; a++) o[a] = e()
        return Lt.makeTensor(o, t, n)
      },
    }),
    Nr = An({
      randomNormal_: function (t, e, n, r, o) {
        if (
          (void 0 === e && (e = 0),
          void 0 === n && (n = 1),
          null != r && 'bool' === r)
        )
          throw new Error('Unsupported data type ' + r)
        for (
          var a = new lr(e, n, r, !1, o), i = dr(t, r), s = 0;
          s < i.values.length;
          s++
        )
          i.values[s] = a.nextValue()
        return i.toTensor()
      },
    }),
    Fr = An({
      randomGamma_: function (t, e, n, r, o) {
        if (
          (void 0 === n && (n = 1),
          void 0 === r && (r = 'float32'),
          null == n && (n = 1),
          null == r && (r = 'float32'),
          'float32' !== r && 'int32' !== r)
        )
          throw new Error('Unsupported data type ' + r)
        for (
          var a = new hr(e, n, r, o), i = dr(t, r), s = 0;
          s < i.values.length;
          s++
        )
          i.values[s] = a.nextValue()
        return i.toTensor()
      },
    }),
    _r = An({
      randomUniform_: function (t, e, n, r, o) {
        void 0 === e && (e = 0),
          void 0 === n && (n = 1),
          void 0 === r && (r = 'float32')
        for (
          var a = dr(t, r), i = new fr(e, n, null, o), s = 0;
          s < a.values.length;
          s++
        )
          a.values[s] = i.nextValue()
        return a.toTensor()
      },
    }),
    Or = An({
      reshape_: function (t, e) {
        var n = mn(t, 'x', 'reshape', null)
        ;(e = _(e, n.size)),
          C(n.size === k(e), function () {
            return 'new shape and old shape must have the same number of elements.'
          })
        var r = { shape: e }
        return Lt.runKernelFunc(
          function (t) {
            return t.reshape(n, e)
          },
          { x: n },
          function (t) {
            return {
              x: function () {
                return t.reshape(n.shape)
              },
            }
          },
          'Reshape',
          r
        )
      },
    }),
    Mr = An({
      spaceToBatchND_: function (t, e, n) {
        var r = mn(t, 'x', 'spaceToBatchND')
        return (
          C(r.rank >= 1 + e.length, function () {
            return (
              'input rank ' +
              r.rank +
              ' should be > than [blockShape] ' +
              e.length
            )
          }),
          C(n.length === e.length, function () {
            return (
              'paddings.shape[0] ' +
              n.length +
              ' must be equal to [blockShape] ' +
              e.length
            )
          }),
          C(
            r.shape.reduce(function (t, r, o) {
              return o > 0 && o <= e.length
                ? t && (r + n[o - 1][0] + n[o - 1][1]) % e[o - 1] == 0
                : t
            }, !0),
            function () {
              return (
                'input spatial dimensions ' +
                r.shape.slice(1) +
                ' with paddings ' +
                n.toString() +
                ' must be divisible by blockShapes ' +
                e.toString()
              )
            }
          ),
          Lt.runKernelFunc(
            function (t) {
              return t.spaceToBatchND(r, e, n)
            },
            { $x: r },
            function (t) {
              return {
                $x: function () {
                  return t.batchToSpaceND(e, n)
                },
              }
            }
          )
        )
      },
    }),
    Br = An({
      squeeze_: function (t, e) {
        var n = mn(t, 'x', 'squeeze')
        return Or(n, M(n.shape, e).newShape)
      },
    }),
    Pr = An({
      stack_: function (t, e) {
        void 0 === e && (e = 0)
        var n = gn(t, 'tensors', 'stack')
        if (
          (C(n.length >= 1, function () {
            return 'Pass at least one tensor to tf.stack'
          }),
          1 === n.length)
        )
          return n[0].expandDims(e)
        var r = n[0].rank,
          o = n[0].shape,
          a = n[0].dtype
        C(e <= r, function () {
          return 'Axis must be <= rank of the tensor'
        }),
          n.forEach(function (t) {
            E(
              o,
              t.shape,
              'All tensors passed to stack must have matching shapes'
            )
          }),
          n.forEach(function (t) {
            C(a === t.dtype, function () {
              return 'All tensors passed to stack must have matching dtypes'
            })
          })
        var i = n.map(function (t) {
          return t.expandDims(e)
        })
        return Yn(i, e)
      },
    }),
    Lr = An({
      tile_: function (t, e) {
        var n = mn(t, 'x', 'tile', null)
        C(n.rank === e.length, function () {
          return (
            'Error in transpose: rank of input ' +
            n.rank +
            ' must match length of reps ' +
            e +
            '.'
          )
        })
        var r = [n],
          o = { reps: e }
        return Lt.runKernelFunc(
          function (t, r) {
            var o = t.tile(n, e)
            return r([n]), o
          },
          { x: n },
          function (t, n) {
            var r = n[0]
            return {
              x: function () {
                var n = Xn(r)
                if (1 === r.rank)
                  for (var o = 0; o < e[0]; ++o)
                    n = n.add(t.slice([o * r.shape[0]], [r.shape[0]]))
                else if (2 === r.rank)
                  for (o = 0; o < e[0]; ++o)
                    for (var a = 0; a < e[1]; ++a)
                      n = n.add(
                        t.slice(
                          [o * r.shape[0], a * r.shape[1]],
                          [r.shape[0], r.shape[1]]
                        )
                      )
                else if (3 === r.rank)
                  for (o = 0; o < e[0]; ++o)
                    for (a = 0; a < e[1]; ++a)
                      for (var i = 0; i < e[2]; ++i)
                        n = n.add(
                          t.slice(
                            [o * r.shape[0], a * r.shape[1], i * r.shape[2]],
                            [r.shape[0], r.shape[1], r.shape[2]]
                          )
                        )
                else {
                  if (4 !== r.rank)
                    throw new Error(
                      'Gradient for tile operation is not implemented for rank-' +
                        r.rank +
                        ' tensors yet.'
                    )
                  for (o = 0; o < e[0]; ++o)
                    for (a = 0; a < e[1]; ++a)
                      for (i = 0; i < e[2]; ++i)
                        for (var s = 0; s < e[3]; ++s)
                          n = n.add(
                            t.slice(
                              [
                                o * r.shape[0],
                                a * r.shape[1],
                                i * r.shape[2],
                                s * r.shape[3],
                              ],
                              [r.shape[0], r.shape[1], r.shape[2], r.shape[3]]
                            )
                          )
                }
                return n
              },
            }
          },
          'Tile',
          o,
          r
        )
      },
    }),
    Wr = An({
      truncatedNormal_: function (t, e, n, r, o) {
        if (
          (void 0 === e && (e = 0),
          void 0 === n && (n = 1),
          null != r && 'bool' === r)
        )
          throw new Error('Unsupported data type ' + r)
        for (
          var a = new lr(e, n, r, !0, o), i = dr(t, r), s = 0;
          s < i.values.length;
          s++
        )
          i.values[s] = a.nextValue()
        return i.toTensor()
      },
    }),
    Ur = An({
      unstack_: function (t, e) {
        void 0 === e && (e = 0), (e = e || 0)
        var n = mn(t, 'x', 'unstack')
        C(e >= -n.shape.length && e < n.shape.length, function () {
          return (
            'Axis = ' +
            e +
            ' is not in [-' +
            n.shape.length +
            ', ' +
            n.shape.length +
            ')'
          )
        }),
          e < 0 && (e += n.shape.length)
        var r = { axis: e }
        return Lt.runKernelFunc(
          function (t) {
            return t.unstack(n, e)
          },
          { x: n },
          function (t) {
            return {
              x: function () {
                return Pr(t, e)
              },
            }
          },
          'Unpack',
          r
        )
      },
    }),
    Vr = function (t, e) {
      return n(this, void 0, void 0, function () {
        var n, o, a, i, s, u, c, l, h, f
        return r(this, function (r) {
          switch (r.label) {
            case 0:
              return (
                (n = mn(t, 'x', 'setdiff1d')),
                (o = mn(e, 'y', 'setdiff1d')),
                C(n.dtype === o.dtype, function () {
                  return (
                    'x and y should have the same dtype, but got x (' +
                    n.dtype +
                    ') and y (' +
                    o.dtype +
                    ').'
                  )
                }),
                C(1 === n.rank, function () {
                  return 'x should be 1D tensor, but got x (' + n.shape + ').'
                }),
                C(1 === o.rank, function () {
                  return 'y should be 1D tensor, but got y (' + o.shape + ').'
                }),
                [4, n.data()]
              )
            case 1:
              return (a = r.sent()), [4, o.data()]
            case 2:
              for (
                i = r.sent(), s = new Set(i), u = 0, h = 0;
                h < a.length;
                h++
              )
                s.has(a[h]) || u++
              for (
                c = new gt([u], n.dtype),
                  l = new gt([u], 'int32'),
                  h = 0,
                  f = 0;
                h < a.length;
                h++
              )
                s.has(a[h]) || ((c.values[f] = a[h]), (l.values[f] = h), f++)
              return [2, [c.toTensor(), l.toTensor()]]
          }
        })
      })
    }
  function zr(t, e, n, r) {
    void 0 === r && (r = !0)
    var o = []
    if (r) (o = o.concat(e.slice(0))).push(t[0] / n), (o = o.concat(t.slice(1)))
    else {
      o = o.concat(t[0])
      for (var a = e.length, i = 0; i < a; ++i)
        o = o.concat([t[i + 1] / e[i], e[i]])
      o = o.concat(t.slice(a + 1))
    }
    return o
  }
  function Gr(t, e, n) {
    void 0 === n && (n = !0)
    var r = []
    if (n) {
      r.push(e)
      for (var o = e + 1; o < t; ++o)
        o <= 2 * e ? (r.push(o), r.push(o - (e + 1))) : r.push(o)
    } else {
      var a = [],
        i = []
      for (o = 1; o < t; ++o)
        o >= 2 * e + 1 || o % 2 == 1 ? i.push(o) : a.push(o)
      r.push.apply(r, a), r.push(0), r.push.apply(r, i)
    }
    return r
  }
  function Hr(t, e, n, r) {
    void 0 === r && (r = !0)
    var o = []
    r ? o.push(t[0] / n) : o.push(t[0] * n)
    for (var a = 1; a < t.length; ++a)
      a <= e.length
        ? r
          ? o.push(e[a - 1] * t[a])
          : o.push(t[a] / e[a - 1])
        : o.push(t[a])
    return o
  }
  function qr(t, e) {
    for (var n = [0], r = 0; r < e; ++r) n.push(t[r][0])
    return n
  }
  function Kr(t, e, n) {
    for (var r = t.slice(0, 1), o = 0; o < n; ++o)
      r.push(t[o + 1] - e[o][0] - e[o][1])
    return r
  }
  function jr(t, e) {
    if (t.rank < 1)
      throw new Error(
        'tf.gatherND() expects the input to be rank 1 or higher, but the rank was ' +
          t.rank +
          '.'
      )
    if (e.rank < 1)
      throw new Error(
        'tf.gatherND() expects the indices to be rank 1 or higher, but the rank was ' +
          e.rank +
          '.'
      )
    if ('int32' !== e.dtype)
      throw new Error(
        'tf.gatherND() expects the indices to be int32 type, but the dtype was ' +
          e.dtype +
          '.'
      )
    if (e.shape[e.rank - 1] > t.rank)
      throw new Error(
        'index innermost dimension length must be <= tensor rank; saw: ' +
          e.shape[e.rank - 1] +
          ' vs. ' +
          t.rank
      )
    if (0 === t.size)
      throw new Error(
        'Requested more than 0 entries, but input is empty. Input shape: ' +
          t.shape +
          '.'
      )
    for (
      var n = e.shape, r = n[n.length - 1], o = 1, a = 0;
      a < n.length - 1;
      ++a
    )
      o *= n[a]
    var i = t.shape,
      s = n.slice()
    s.pop()
    var u = 1
    for (a = r; a < t.rank; ++a) (u *= i[a]), s.push(i[a])
    var c = $(t.shape)
      .map(function (t) {
        return t / u
      })
      .concat([1])
      .slice(0, r)
    return [s, o, u, c]
  }
  var Xr = Object.freeze({ prepareAndValidate: jr }),
    Yr = 30
  function $r(t) {
    return t <= Yr ? t : Y(t, Math.floor(Math.sqrt(t)))
  }
  function Qr(t, e, n) {
    var r = e.rank > 1 ? e.shape[e.rank - 1] : 1,
      o = e.rank > 1 ? e.rank - 1 : 1,
      a =
        'Must have updates.shape = indices.shape[:batchDim] + shape[sliceDim:], got updates.shape: ' +
        n.shape +
        ', indices.shape: ' +
        e.shape +
        ', shape: ' +
        t +
        ', sliceDim: ' +
        r +
        ', and batchDim: ' +
        o +
        '.'
    if (n.rank < o) throw new Error(a + ' update.rank < ' + o + '. ')
    if (t.length < r + (n.rank - o))
      throw new Error(a + ' Output shape length < ' + (r + (n.rank - o)))
    if (n.rank !== o + t.length - r)
      throw new Error(a + ' update.rank != ' + (o + t.length - r))
    for (var i = 0; i < o; ++i)
      if (n.shape[i] !== e.shape[i])
        throw new Error(
          a +
            ' updates.shape[' +
            i +
            '] (' +
            n.shape[i] +
            ') != indices.shape[' +
            i +
            '] (' +
            e.shape[i] +
            ').'
        )
    for (i = 0; i < n.rank - o; ++i)
      if (n.shape[i + o] !== t[i + r])
        throw new Error(
          a +
            ' updates.shape[' +
            (i + o) +
            '] (' +
            n.shape[i + o] +
            ') != shape[' +
            (i + o) +
            '] (' +
            t[i + o] +
            ')'
        )
  }
  function Jr(t, e, n) {
    if (e.rank < 1)
      throw new Error(
        'tf.scatterND() expects the indices to be rank 1 or higher, but the rank was ' +
          e.rank +
          '.'
      )
    if (t.rank < 1)
      throw new Error(
        'tf.scatterND() expects the updates to be rank 1 or higher, but the rank was ' +
          t.rank +
          '.'
      )
    if ('int32' !== e.dtype)
      throw new Error(
        "The dtype of 'indices' should be int32, but got dtype: " + e.dtype
      )
    if (n.length < 1)
      throw new Error(
        'Output rank must be greater or equal to 1, but got shape: ' + n
      )
    if (0 === n.length) {
      if (0 === e.size)
        throw new Error(
          'Indices specified for empty output. indices shape: ' + e.shape
        )
      if (0 === t.size)
        throw new Error(
          'Updates specified for empty output. updates shape: ' + t.shape
        )
    }
    Qr(n, e, t)
  }
  function Zr(t, e, n) {
    for (
      var r = e.shape.length,
        o = r > 1 ? e.shape[r - 1] : 1,
        a = n.length,
        i = 1,
        s = o;
      s < a;
      ++s
    )
      i *= n[s]
    var u = o < 1 ? 1 : o
    return {
      sliceRank: o,
      numUpdates: k(e.shape) / u,
      sliceSize: i,
      strides: $(n.slice(0, o)).concat([1]),
      outputSize: k(n),
    }
  }
  var to = Object.freeze({
    validateUpdateShape: Qr,
    validateInput: Jr,
    calculateShapes: Zr,
  })
  function eo(t, e, n) {
    C(t.rank === e.length, function () {
      return (
        'Error in slice' +
        t.rank +
        'D: Length of begin ' +
        e +
        ' must match the rank of the array (' +
        t.rank +
        ').'
      )
    }),
      C(t.rank === n.length, function () {
        return (
          'Error in slice' +
          t.rank +
          'D: Length of size ' +
          n +
          ' must match the rank of the array (' +
          t.rank +
          ').'
        )
      })
    for (
      var r = function (r) {
          C(e[r] + n[r] <= t.shape[r], function () {
            return (
              'Error in slice' +
              t.rank +
              'D: begin[' +
              r +
              '] + size[' +
              r +
              '] (' +
              (e[r] + n[r]) +
              ') would overflow input.shape[' +
              r +
              '] (' +
              t.shape[r] +
              ')'
            )
          })
        },
        o = 0;
      o < t.rank;
      ++o
    )
      r(o)
  }
  function no(t) {
    for (var e = [], n = 0; t > 0; ) 1 & t && e.push(n), (t /= 2), n++
    return e
  }
  function ro(t, e, n) {
    for (var r = [], o = 0; o < t.length; o++)
      r[o] = Math.ceil((e[o] - t[o]) / n[o])
    return r
  }
  function oo(t, e, n, r, o) {
    var a = e[o],
      i = n[o] || 1
    ;(t & (1 << o) || null == a) &&
      (a = i > 0 ? Number.MIN_SAFE_INTEGER : Number.MAX_SAFE_INTEGER)
    var s = r[o]
    return a < 0 && (a += s), (a = x(0, a, s - 1))
  }
  function ao(t, e, n, r, o) {
    var a = e[o],
      i = n[o] || 1
    ;(t & (1 << o) || null == a) &&
      (a = i > 0 ? Number.MAX_SAFE_INTEGER : Number.MIN_SAFE_INTEGER)
    var s = r[o]
    return a < 0 && (a += s), (a = i > 0 ? x(0, a, s) : x(-1, a, s - 1))
  }
  function io(t, e, n) {
    for (var r = n.length, o = 0; o < n.length; o++)
      if (n[o] > 1) {
        r = o
        break
      }
    for (o = r + 1; o < n.length; o++) if (e[o] > 0 || n[o] !== t[o]) return !1
    return !0
  }
  function so(t, e) {
    for (
      var n = t.length > 0 ? t[t.length - 1] : 1, r = 0;
      r < t.length - 1;
      r++
    )
      n += t[r] * e[r]
    return n
  }
  var uo = Object.freeze({
    assertParamsValid: eo,
    maskToAxes: no,
    computeOutShape: ro,
    startForAxis: oo,
    stopForAxis: ao,
    isSliceContinous: io,
    computeFlatOffset: so,
  })
  function co(t) {
    return (
      C(X(t), function () {
        return 'The f passed in grad(f) must be a function'
      }),
      function (e, n) {
        var r = mn(e, 'x', 'tf.grad', null),
          o = null != n ? mn(n, 'dy', 'tf.grad') : null
        return Lt.tidy(function () {
          var e = Lt.gradients(
              function () {
                return t(r)
              },
              [r],
              o
            ),
            n = e.value,
            a = e.grads
          return (
            null != o &&
              E(
                n.shape,
                o.shape,
                'The shape of dy passed in grad(f)(x, dy) must match the shape returned by f(x)'
              ),
            mo(a),
            a[0]
          )
        })
      }
    )
  }
  function lo(t) {
    return (
      C(X(t), function () {
        return 'The f passed in grads(f) must be a function'
      }),
      function (e, n) {
        C(Array.isArray(e), function () {
          return 'The args passed in grads(f)(args) must be an array of `Tensor`s or `TensorLike`s'
        })
        var r = gn(e, 'args', 'tf.grads', null),
          o = null != n ? mn(n, 'dy', 'tf.grads') : null
        return Lt.tidy(function () {
          var e = Lt.gradients(
              function () {
                return t.apply(void 0, r)
              },
              r,
              o
            ),
            n = e.value,
            a = e.grads
          return (
            null != o &&
              E(
                n.shape,
                o.shape,
                'The shape of dy passed in grads(f)([x1,...], dy) must match the shape returned by f([x1,...])'
              ),
            mo(a),
            a
          )
        })
      }
    )
  }
  function ho(t) {
    return (
      C(X(t), function () {
        return 'The f passed in valueAndGrad(f) must be a function'
      }),
      function (e, n) {
        C(e instanceof wt, function () {
          return 'The x passed in valueAndGrad(f)(x) must be a tensor'
        }),
          C(null == n || n instanceof wt, function () {
            return 'The dy passed in valueAndGrad(f)(x, dy) must be a tensor'
          })
        var r = Lt.gradients(
            function () {
              return t(e)
            },
            [e],
            n
          ),
          o = r.grads,
          a = r.value
        return mo(o), { grad: o[0], value: a }
      }
    )
  }
  function fo(t) {
    return (
      C(X(t), function () {
        return 'The f passed in valueAndGrads(f) must be a function'
      }),
      function (e, n) {
        C(
          Array.isArray(e) &&
            e.every(function (t) {
              return t instanceof wt
            }),
          function () {
            return 'The args passed in valueAndGrads(f)(args) must be array of tensors'
          }
        ),
          C(null == n || n instanceof wt, function () {
            return 'The dy passed in valueAndGrads(f)(args, dy) must be a tensor'
          })
        var r = Lt.gradients(
          function () {
            return t.apply(void 0, e)
          },
          e,
          n
        )
        return (
          null != n &&
            E(
              r.value.shape,
              n.shape,
              'The shape of dy passed in valueAndGrads(f)([x1,...], dy) must match the shape returned by f([x1,...])'
            ),
          mo(r.grads),
          r
        )
      }
    )
  }
  function po(t, e) {
    C(X(t), function () {
      return 'The f passed in variableGrads(f) must be a function'
    }),
      C(
        null == e ||
          (Array.isArray(e) &&
            e.every(function (t) {
              return t instanceof St
            })),
        function () {
          return 'The varList passed in variableGrads(f, varList) must be an array of variables'
        }
      )
    var n = null != e
    if (!n)
      for (var r in ((e = []), Lt.registeredVariables))
        e.push(Lt.registeredVariables[r])
    var o = n
        ? e.filter(function (t) {
            return !t.trainable
          })
        : null,
      a = e.length
    C(
      (e = e.filter(function (t) {
        return t.trainable
      })).length > 0,
      function () {
        return (
          'variableGrads() expects at least one of the input variables to be trainable, but none of the ' +
          a +
          ' variables is trainable.'
        )
      }
    )
    var i = Lt.gradients(t, e, null, !0),
      s = i.value,
      u = i.grads
    C(
      u.some(function (t) {
        return null != t
      }),
      function () {
        return 'Cannot find a connection between any variable and the result of the loss function y=f(x). Please make sure the operations that use variables are inside the function f passed to minimize().'
      }
    ),
      C(0 === s.rank, function () {
        return (
          'The f passed in variableGrads(f) must return a scalar, but it returned a rank-' +
          s.rank +
          ' tensor'
        )
      })
    var c = {}
    return (
      e.forEach(function (t, e) {
        null != u[e] && (c[t.name] = u[e])
      }),
      null != o &&
        o.forEach(function (t) {
          return (c[t.name] = null)
        }),
      { value: s, grads: c }
    )
  }
  function vo(t) {
    return Lt.customGrad(t)
  }
  function mo(t) {
    if (
      t.filter(function (t) {
        return null == t
      }).length > 0
    )
      throw new Error(
        'Cannot compute gradient of y=f(x) with respect to x. Make sure that\n    the f you passed encloses all operations that lead from x to y.'
      )
  }
  var go = An({
      softmax_: function (t, e) {
        void 0 === e && (e = -1)
        var n = mn(t, 'logits', 'softmax', 'float32')
        if ((-1 === e && (e = n.rank - 1), e !== n.rank - 1))
          throw Error(
            'Softmax along a non-last dimension is not yet supported. Logits was rank ' +
              n.rank +
              ' and dim was ' +
              e
          )
        return Lt.runKernelFunc(
          function (t, r) {
            var o = t.softmax(n, e)
            return r([o]), o
          },
          { logits: n },
          function (t, n) {
            var r = n[0],
              o = t.mul(r)
            return {
              logits: function () {
                return o.sub(o.sum([e], !0).mul(r))
              },
            }
          },
          'Softmax',
          { dim: e },
          [],
          [!0]
        )
      },
    }),
    yo = An({
      logSoftmax_: function (t, e) {
        void 0 === e && (e = -1)
        var n = mn(t, 'logits', 'logSoftmax')
        if ((-1 === e && (e = n.rank - 1), e !== n.rank - 1))
          throw Error(
            'Log Softmax along a non-last dimension is not yet supported. Logits was rank ' +
              n.rank +
              ' and axis was ' +
              e
          )
        return vo(function (t, n) {
          var r = t.max(e, !0),
            o = t.sub(r),
            a = o.toFloat().sub(o.exp().sum(e, !0).log())
          n([a])
          return {
            value: a,
            gradFunc: function (t, n) {
              var r = n[0].exp()
              return t.sub(t.sum(e, !0).mul(r))
            },
          }
        })(n)
      },
    }),
    xo = (function () {
      function t(t, e) {
        ;(this.backend = t),
          (this.dataMover = e),
          (this.data = new WeakMap()),
          (this.dataIdsCount = 0)
      }
      return (
        (t.prototype.get = function (t) {
          return (
            this.data.has(t) || this.dataMover.moveData(this.backend, t),
            this.data.get(t)
          )
        }),
        (t.prototype.set = function (t, e) {
          this.dataIdsCount++, this.data.set(t, e)
        }),
        (t.prototype.has = function (t) {
          return this.data.has(t)
        }),
        (t.prototype.delete = function (t) {
          return this.dataIdsCount--, this.data.delete(t)
        }),
        (t.prototype.numDataIds = function () {
          return this.dataIdsCount
        }),
        t
      )
    })(),
    bo = (function () {
      function t() {}
      return (
        (t.prototype.time = function (t) {
          return wo('time')
        }),
        (t.prototype.read = function (t) {
          return wo('read')
        }),
        (t.prototype.readSync = function (t) {
          return wo('readSync')
        }),
        (t.prototype.numDataIds = function () {
          return wo('numDataIds')
        }),
        (t.prototype.disposeData = function (t) {
          return wo('disposeData')
        }),
        (t.prototype.write = function (t, e, n) {
          return wo('write')
        }),
        (t.prototype.move = function (t, e, n, r) {
          return wo('move')
        }),
        (t.prototype.memory = function () {
          return wo('memory')
        }),
        (t.prototype.floatPrecision = function () {
          return wo('floatPrecision')
        }),
        (t.prototype.epsilon = function () {
          return 32 === this.floatPrecision() ? 1e-7 : 1e-4
        }),
        (t.prototype.batchMatMul = function (t, e, n, r) {
          return wo('batchMatMul')
        }),
        (t.prototype.fusedBatchMatMul = function (t) {
          t.a,
            t.b,
            t.transposeA,
            t.transposeB,
            t.bias,
            t.activation,
            t.preluActivationWeights
          return wo('fusedBatchMatMul')
        }),
        (t.prototype.slice = function (t, e, n) {
          return wo('slice')
        }),
        (t.prototype.stridedSlice = function (t, e, n, r) {
          return wo('stridedSlice')
        }),
        (t.prototype.unstack = function (t, e) {
          return wo('unstack')
        }),
        (t.prototype.reverse = function (t, e) {
          return wo('reverse')
        }),
        (t.prototype.concat = function (t, e) {
          return wo('concat')
        }),
        (t.prototype.neg = function (t) {
          return wo('neg')
        }),
        (t.prototype.add = function (t, e) {
          return wo('add')
        }),
        (t.prototype.addN = function (t) {
          return wo('addN')
        }),
        (t.prototype.subtract = function (t, e) {
          return wo('subtract')
        }),
        (t.prototype.multiply = function (t, e) {
          return wo('multiply')
        }),
        (t.prototype.realDivide = function (t, e) {
          return wo('realDivide')
        }),
        (t.prototype.floorDiv = function (t, e) {
          return wo('floorDiv')
        }),
        (t.prototype.sum = function (t, e) {
          return wo('sum')
        }),
        (t.prototype.prod = function (t, e) {
          return wo('prod')
        }),
        (t.prototype.unsortedSegmentSum = function (t, e, n) {
          return wo('unsortedSegmentSum')
        }),
        (t.prototype.argMin = function (t, e) {
          return wo('argMin')
        }),
        (t.prototype.argMax = function (t, e) {
          return wo('argMax')
        }),
        (t.prototype.equal = function (t, e) {
          return wo('equal')
        }),
        (t.prototype.notEqual = function (t, e) {
          return wo('notEqual')
        }),
        (t.prototype.less = function (t, e) {
          return wo('less')
        }),
        (t.prototype.lessEqual = function (t, e) {
          return wo('lessEqual')
        }),
        (t.prototype.greater = function (t, e) {
          return wo('greater')
        }),
        (t.prototype.greaterEqual = function (t, e) {
          return wo('greaterEqual')
        }),
        (t.prototype.logicalNot = function (t) {
          return wo('logicalNot')
        }),
        (t.prototype.logicalAnd = function (t, e) {
          return wo('logicalAnd')
        }),
        (t.prototype.logicalOr = function (t, e) {
          return wo('logicalOr')
        }),
        (t.prototype.where = function (t) {
          return wo('where')
        }),
        (t.prototype.select = function (t, e, n) {
          return wo('select')
        }),
        (t.prototype.topk = function (t, e, n) {
          return wo('topk')
        }),
        (t.prototype.min = function (t, e) {
          return wo('min')
        }),
        (t.prototype.minimum = function (t, e) {
          return wo('minimum')
        }),
        (t.prototype.mod = function (t, e) {
          return wo('mod')
        }),
        (t.prototype.max = function (t, e) {
          return wo('max')
        }),
        (t.prototype.maximum = function (t, e) {
          return wo('maximum')
        }),
        (t.prototype.all = function (t, e) {
          return wo('all')
        }),
        (t.prototype.any = function (t, e) {
          return wo('any')
        }),
        (t.prototype.squaredDifference = function (t, e) {
          return wo('squaredDifference')
        }),
        (t.prototype.ceil = function (t) {
          return wo('ceil')
        }),
        (t.prototype.floor = function (t) {
          return wo('floor')
        }),
        (t.prototype.round = function (t) {
          return wo('round')
        }),
        (t.prototype.sign = function (t) {
          return wo('sign')
        }),
        (t.prototype.isNaN = function (t) {
          return wo('isNaN')
        }),
        (t.prototype.isInf = function (t) {
          return wo('isInf')
        }),
        (t.prototype.isFinite = function (t) {
          return wo('isFinite')
        }),
        (t.prototype.pow = function (t, e) {
          return wo('pow')
        }),
        (t.prototype.exp = function (t) {
          return wo('exp')
        }),
        (t.prototype.expm1 = function (t) {
          return wo('expm1')
        }),
        (t.prototype.softmax = function (t, e) {
          return wo('softmax')
        }),
        (t.prototype.log = function (t) {
          return wo('log')
        }),
        (t.prototype.log1p = function (t) {
          return wo('log1p')
        }),
        (t.prototype.sqrt = function (t) {
          return wo('sqrt')
        }),
        (t.prototype.rsqrt = function (t) {
          return wo('rsqrt')
        }),
        (t.prototype.square = function (t) {
          return wo('square')
        }),
        (t.prototype.reciprocal = function (t) {
          return wo('reciprocal')
        }),
        (t.prototype.relu = function (t) {
          return wo('relu')
        }),
        (t.prototype.relu6 = function (t) {
          return wo('relu6')
        }),
        (t.prototype.prelu = function (t, e) {
          return wo('prelu')
        }),
        (t.prototype.elu = function (t) {
          return wo('elu')
        }),
        (t.prototype.eluDer = function (t, e) {
          return wo('eluDer')
        }),
        (t.prototype.selu = function (t) {
          return wo('selu')
        }),
        (t.prototype.int = function (t) {
          return wo('int')
        }),
        (t.prototype.clip = function (t, e, n) {
          return wo('clip')
        }),
        (t.prototype.abs = function (t) {
          return wo('abs')
        }),
        (t.prototype.complexAbs = function (t) {
          return wo('complexAbs')
        }),
        (t.prototype.sigmoid = function (t) {
          return wo('sigmoid')
        }),
        (t.prototype.softplus = function (t) {
          return wo('softplus')
        }),
        (t.prototype.sin = function (t) {
          return wo('sin')
        }),
        (t.prototype.cos = function (t) {
          return wo('cos')
        }),
        (t.prototype.tan = function (t) {
          return wo('tan')
        }),
        (t.prototype.asin = function (t) {
          return wo('asin')
        }),
        (t.prototype.acos = function (t) {
          return wo('acos')
        }),
        (t.prototype.atan = function (t) {
          return wo('atan')
        }),
        (t.prototype.atan2 = function (t, e) {
          return wo('atan2')
        }),
        (t.prototype.sinh = function (t) {
          return wo('sinh')
        }),
        (t.prototype.cosh = function (t) {
          return wo('cosh')
        }),
        (t.prototype.tanh = function (t) {
          return wo('tanh')
        }),
        (t.prototype.asinh = function (t) {
          return wo('asinh')
        }),
        (t.prototype.acosh = function (t) {
          return wo('acosh')
        }),
        (t.prototype.atanh = function (t) {
          return wo('atanh')
        }),
        (t.prototype.erf = function (t) {
          return wo('erf')
        }),
        (t.prototype.step = function (t, e) {
          return wo('step')
        }),
        (t.prototype.fusedConv2d = function (t) {
          t.input,
            t.filter,
            t.convInfo,
            t.bias,
            t.activation,
            t.preluActivationWeights
          return wo('fusedConv2d')
        }),
        (t.prototype.conv2d = function (t, e, n) {
          return wo('conv2d')
        }),
        (t.prototype.conv2dDerInput = function (t, e, n) {
          return wo('conv2dDerInput')
        }),
        (t.prototype.conv2dDerFilter = function (t, e, n) {
          return wo('conv2dDerFilter')
        }),
        (t.prototype.fusedDepthwiseConv2D = function (t) {
          t.input,
            t.filter,
            t.convInfo,
            t.bias,
            t.activation,
            t.preluActivationWeights
          return wo('fusedDepthwiseConv2D')
        }),
        (t.prototype.depthwiseConv2D = function (t, e, n) {
          return wo('depthwiseConv2D')
        }),
        (t.prototype.depthwiseConv2DDerInput = function (t, e, n) {
          return wo('depthwiseConv2DDerInput')
        }),
        (t.prototype.depthwiseConv2DDerFilter = function (t, e, n) {
          return wo('depthwiseConv2DDerFilter')
        }),
        (t.prototype.conv3d = function (t, e, n) {
          return wo('conv3d')
        }),
        (t.prototype.conv3dDerInput = function (t, e, n) {
          return wo('conv3dDerInput')
        }),
        (t.prototype.conv3dDerFilter = function (t, e, n) {
          return wo('conv3dDerFilter')
        }),
        (t.prototype.maxPool = function (t, e) {
          return wo('maxPool')
        }),
        (t.prototype.maxPoolBackprop = function (t, e, n, r) {
          return wo('maxPoolBackprop')
        }),
        (t.prototype.avgPool = function (t, e) {
          return wo('avgPool')
        }),
        (t.prototype.avgPoolBackprop = function (t, e, n) {
          return wo('avgPoolBackprop')
        }),
        (t.prototype.avgPool3d = function (t, e) {
          return wo('avgPool3d')
        }),
        (t.prototype.avgPool3dBackprop = function (t, e, n) {
          return wo('avgPool3dBackprop')
        }),
        (t.prototype.maxPool3d = function (t, e) {
          return wo('maxPool3d')
        }),
        (t.prototype.maxPool3dBackprop = function (t, e, n, r) {
          return wo('maxPool3dBackprop')
        }),
        (t.prototype.reshape = function (t, e) {
          return wo('reshape')
        }),
        (t.prototype.cast = function (t, e) {
          return wo('cast')
        }),
        (t.prototype.tile = function (t, e) {
          return wo('tile')
        }),
        (t.prototype.pad = function (t, e, n) {
          return wo('pad')
        }),
        (t.prototype.transpose = function (t, e) {
          return wo('transpose')
        }),
        (t.prototype.gather = function (t, e, n) {
          return wo('gather')
        }),
        (t.prototype.gatherND = function (t, e) {
          return wo('gatherND')
        }),
        (t.prototype.scatterND = function (t, e, n) {
          return wo('scatterND')
        }),
        (t.prototype.batchToSpaceND = function (t, e, n) {
          return wo('batchToSpaceND')
        }),
        (t.prototype.spaceToBatchND = function (t, e, n) {
          return wo('spaceToBatchND')
        }),
        (t.prototype.resizeBilinear = function (t, e, n, r) {
          return wo('resizeBilinear')
        }),
        (t.prototype.resizeBilinearBackprop = function (t, e, n) {
          return wo('resizeBilinearBackprop')
        }),
        (t.prototype.resizeNearestNeighbor = function (t, e, n, r) {
          return wo('resizeNearestNeighbor')
        }),
        (t.prototype.resizeNearestNeighborBackprop = function (t, e, n) {
          return wo('resizeNearestNeighborBackprop')
        }),
        (t.prototype.batchNormalization = function (t, e, n, r, o, a) {
          return wo('batchNormalization')
        }),
        (t.prototype.localResponseNormalization4D = function (t, e, n, r, o) {
          return wo('localResponseNormalization4D')
        }),
        (t.prototype.LRNGrad = function (t, e, n, r, o, a, i) {
          return wo('LRNGrad')
        }),
        (t.prototype.multinomial = function (t, e, n, r) {
          return wo('multinomial')
        }),
        (t.prototype.oneHot = function (t, e, n, r) {
          return wo('oneHot')
        }),
        (t.prototype.cumsum = function (t, e, n, r) {
          return wo('cumsum')
        }),
        (t.prototype.nonMaxSuppression = function (t, e, n, r, o) {
          return wo('nonMaxSuppression')
        }),
        (t.prototype.fft = function (t) {
          return wo('fft')
        }),
        (t.prototype.ifft = function (t) {
          return wo('ifft')
        }),
        (t.prototype.complex = function (t, e) {
          return wo('complex')
        }),
        (t.prototype.real = function (t) {
          return wo('real')
        }),
        (t.prototype.imag = function (t) {
          return wo('imag')
        }),
        (t.prototype.cropAndResize = function (t, e, n, r, o, a) {
          return wo('cropAndResize')
        }),
        (t.prototype.depthToSpace = function (t, e, n) {
          return wo('depthToSpace')
        }),
        (t.prototype.split = function (t, e, n) {
          return wo('split')
        }),
        (t.prototype.sparseToDense = function (t, e, n, r) {
          return wo('sparseToDense')
        }),
        (t.prototype.diag = function (t) {
          return wo('diag')
        }),
        (t.prototype.fill = function (t, e, n) {
          return wo('fill')
        }),
        (t.prototype.onesLike = function (t) {
          return wo('onesLike')
        }),
        (t.prototype.zerosLike = function (t) {
          return wo('zerosLike')
        }),
        (t.prototype.linspace = function (t, e, n) {
          return wo('linspace')
        }),
        (t.prototype.dispose = function () {
          return wo('dispose')
        }),
        t
      )
    })()
  function wo(t) {
    throw new Error(
      "'" +
        t +
        "' not yet implemented or not found in the registry. Did you forget to import the kernel?"
    )
  }
  function Co(t, e) {
    for (var n = t.length, r = [], o = 0; o < n; o++) {
      var a = n - 1 - o,
        i = t[a] || 1
      ;(e[e.length - 1 - o] || 1) > 1 && 1 === i && r.unshift(a)
    }
    return r
  }
  function Eo(t, e) {
    for (var n = [], r = 0; r < e.length; r++) {
      var o = t[t.length - r - 1],
        a = e.length - r - 1,
        i = e[a]
      ;(null == o || (1 === o && i > 1)) && n.unshift(a)
    }
    return n
  }
  function Ro(t, e) {
    for (var n = [], r = Math.max(t.length, e.length), o = 0; o < r; o++) {
      var a = t[t.length - o - 1]
      null == a && (a = 1)
      var i = e[e.length - o - 1]
      if ((null == i && (i = 1), 1 === a)) n.unshift(i)
      else if (1 === i) n.unshift(a)
      else {
        if (a !== i)
          throw Error(
            'Operands could not be broadcast together with shapes ' +
              t +
              ' and ' +
              e +
              '.'
          )
        n.unshift(a)
      }
    }
    return n
  }
  function Io(t, e, n, r, o, a, i) {
    void 0 === i && (i = 'channelsLast')
    var s,
      u = To(e),
      c = u[0],
      l = u[1]
    if ('channelsLast' === i) s = [c, l, t[3], t[3]]
    else {
      if ('channelsFirst' !== i) throw new Error('Unknown dataFormat ' + i)
      s = [c, l, t[1], t[1]]
    }
    return So(t, s, n, r, o, a, !1, i)
  }
  function ko(t, e, n, r, o, a, i) {
    void 0 === i && (i = 'NDHWC')
    var s,
      u,
      c = No(e),
      l = c[0],
      h = c[1],
      f = c[2]
    if ('NDHWC' === i) (u = 'channelsLast'), (s = [l, h, f, t[4], t[4]])
    else {
      if ('NCDHW' !== i) throw new Error('Unknown dataFormat ' + i)
      ;(u = 'channelsFirst'), (s = [l, h, f, t[1], t[1]])
    }
    return Ao(t, s, n, r, o, !1, u, a)
  }
  function So(t, e, n, r, o, a, i, s) {
    void 0 === i && (i = !1), void 0 === s && (s = 'channelsLast')
    var u = [-1, -1, -1, -1],
      c = u[0],
      l = u[1],
      h = u[2],
      f = u[3]
    if ('channelsLast' === s) (c = t[0]), (l = t[1]), (h = t[2]), (f = t[3])
    else {
      if ('channelsFirst' !== s) throw new Error('Unknown dataFormat ' + s)
      ;(c = t[0]), (f = t[1]), (l = t[2]), (h = t[3])
    }
    var d,
      p = e[0],
      v = e[1],
      m = e[3],
      g = To(n),
      y = g[0],
      x = g[1],
      b = To(r),
      w = b[0],
      E = b[1],
      R = Fo(p, w),
      I = Fo(v, E),
      k = (function (t, e, n, r, o, a, i, s) {
        var u, c, l
        if ('number' == typeof t) {
          u = {
            top: t,
            bottom: t,
            left: t,
            right: t,
            type: 0 === t ? 'VALID' : 'NUMBER',
          }
          var h = (function (t, e, n, r, o) {
            null == r && (r = Do(t, e, n))
            var a = t[0],
              i = t[1],
              s = _o((a - e + 2 * r) / n + 1, o)
            C(A(s), function () {
              return (
                'The output # of rows (' +
                s +
                ') must be an integer. Change the stride and/or zero pad parameters'
              )
            })
            var u = _o((i - e + 2 * r) / n + 1, o)
            return (
              C(A(u), function () {
                return (
                  'The output # of columns (' +
                  u +
                  ') must be an integer. Change the stride and/or zero pad parameters'
                )
              }),
              [s, u]
            )
          })([e, n], a, r, t, s)
          ;(c = h[0]), (l = h[1])
        } else if ('same' === t) {
          ;(c = Math.ceil(e / r)), (l = Math.ceil(n / o))
          var f = Math.max(0, (c - 1) * r + a - e),
            d = Math.max(0, (l - 1) * o + i - n),
            p = Math.floor(f / 2),
            v = f - p,
            m = Math.floor(d / 2)
          u = { top: p, bottom: v, left: m, right: d - m, type: 'SAME' }
        } else {
          if ('valid' !== t) throw Error('Unknown padding parameter: ' + t)
          ;(u = { top: 0, bottom: 0, left: 0, right: 0, type: 'VALID' }),
            (c = Math.ceil((e - a + 1) / r)),
            (l = Math.ceil((n - i + 1) / o))
        }
        return { padInfo: u, outHeight: c, outWidth: l }
      })(o, l, h, y, x, R, I, a),
      S = k.padInfo,
      D = k.outHeight,
      T = k.outWidth,
      N = i ? m * f : m
    return (
      'channelsFirst' === s
        ? (d = [c, N, D, T])
        : 'channelsLast' === s && (d = [c, D, T, N]),
      {
        batchSize: c,
        dataFormat: s,
        inHeight: l,
        inWidth: h,
        inChannels: f,
        outHeight: D,
        outWidth: T,
        outChannels: N,
        padInfo: S,
        strideHeight: y,
        strideWidth: x,
        filterHeight: p,
        filterWidth: v,
        effectiveFilterHeight: R,
        effectiveFilterWidth: I,
        dilationHeight: w,
        dilationWidth: E,
        inShape: t,
        outShape: d,
        filterShape: e,
      }
    )
  }
  function Ao(t, e, n, r, o, a, i, s) {
    void 0 === a && (a = !1), void 0 === i && (i = 'channelsLast')
    var u = [-1, -1, -1, -1, -1],
      c = u[0],
      l = u[1],
      h = u[2],
      f = u[3],
      d = u[4]
    if ('channelsLast' === i)
      (c = t[0]), (l = t[1]), (h = t[2]), (f = t[3]), (d = t[4])
    else {
      if ('channelsFirst' !== i) throw new Error('Unknown dataFormat ' + i)
      ;(c = t[0]), (d = t[1]), (l = t[2]), (h = t[3]), (f = t[4])
    }
    var p,
      v = e[0],
      m = e[1],
      g = e[2],
      y = e[4],
      x = No(n),
      b = x[0],
      w = x[1],
      E = x[2],
      R = No(r),
      I = R[0],
      k = R[1],
      S = R[2],
      D = Fo(v, I),
      T = Fo(m, k),
      N = Fo(g, S),
      F = (function (t, e, n, r, o, a, i, s, u, c, l) {
        var h, f, d, p
        if ('number' == typeof t) {
          h = {
            top: t,
            bottom: t,
            left: t,
            right: t,
            front: t,
            back: t,
            type: 0 === t ? 'VALID' : 'NUMBER',
          }
          var v = (function (t, e, n, r, o, a) {
            null == o && (o = Do(t, e, r))
            var i = t[0],
              s = t[1],
              u = t[2],
              c = _o((i - e + 2 * o) / r + 1, a)
            C(A(c), function () {
              return (
                'The output # of depths (' +
                c +
                ') must be an integer. Change the stride and/or zero pad parameters'
              )
            })
            var l = _o((s - e + 2 * o) / r + 1, a)
            C(A(l), function () {
              return (
                'The output # of rows (' +
                l +
                ') must be an integer. Change the stride and/or zero pad parameters'
              )
            })
            var h = _o((u - e + 2 * o) / r + 1, a)
            return (
              C(A(h), function () {
                return (
                  'The output # of columns (' +
                  h +
                  ') must be an integer. Change the stride and/or zero pad parameters'
                )
              }),
              [c, l, h, n]
            )
          })([e, n, r, 1], s, 1, o, t, l)
          ;(f = v[0]), (d = v[1]), (p = v[2])
        } else if ('same' === t) {
          ;(f = Math.ceil(e / o)),
            (d = Math.ceil(n / a)),
            (p = Math.ceil(r / i))
          var m = (f - 1) * o + s - e,
            g = (d - 1) * a + u - n,
            y = (p - 1) * i + c - r,
            x = Math.floor(m / 2),
            b = m - x,
            w = Math.floor(g / 2),
            E = g - w,
            R = Math.floor(y / 2)
          h = {
            top: w,
            bottom: E,
            left: R,
            right: y - R,
            front: x,
            back: b,
            type: 'SAME',
          }
        } else {
          if ('valid' !== t) throw Error('Unknown padding parameter: ' + t)
          ;(h = {
            top: 0,
            bottom: 0,
            left: 0,
            right: 0,
            front: 0,
            back: 0,
            type: 'VALID',
          }),
            (f = Math.ceil((e - s + 1) / o)),
            (d = Math.ceil((n - u + 1) / a)),
            (p = Math.ceil((r - c + 1) / i))
        }
        return { padInfo: h, outDepth: f, outHeight: d, outWidth: p }
      })(o, l, h, f, b, w, E, D, T, N, s),
      _ = F.padInfo,
      O = F.outDepth,
      M = F.outHeight,
      B = F.outWidth,
      P = a ? y * d : y
    return (
      'channelsFirst' === i
        ? (p = [c, P, O, M, B])
        : 'channelsLast' === i && (p = [c, O, M, B, P]),
      {
        batchSize: c,
        dataFormat: i,
        inDepth: l,
        inHeight: h,
        inWidth: f,
        inChannels: d,
        outDepth: O,
        outHeight: M,
        outWidth: B,
        outChannels: P,
        padInfo: _,
        strideDepth: b,
        strideHeight: w,
        strideWidth: E,
        filterDepth: v,
        filterHeight: m,
        filterWidth: g,
        effectiveFilterDepth: D,
        effectiveFilterHeight: T,
        effectiveFilterWidth: N,
        dilationDepth: I,
        dilationHeight: k,
        dilationWidth: S,
        inShape: t,
        outShape: p,
        filterShape: e,
      }
    )
  }
  function Do(t, e, n, r) {
    void 0 === r && (r = 1)
    var o = Fo(e, r)
    return Math.floor((t[0] * (n - 1) - n + o) / 2)
  }
  function To(t) {
    return 'number' == typeof t
      ? [t, t, t]
      : 2 === t.length
      ? [t[0], t[1], 1]
      : t
  }
  function No(t) {
    return 'number' == typeof t ? [t, t, t] : t
  }
  function Fo(t, e) {
    return e <= 1 ? t : t + (t - 1) * (e - 1)
  }
  function _o(t, e) {
    if (!e) return t
    switch (e) {
      case 'round':
        return Math.round(t)
      case 'ceil':
        return Math.ceil(t)
      case 'floor':
        return Math.floor(t)
      default:
        throw new Error('Unknown roundingMode ' + e)
    }
  }
  function Oo(t) {
    var e = To(t),
      n = e[0],
      r = e[1],
      o = e[2]
    return 1 === n && 1 === r && 1 === o
  }
  function Mo(t, e) {
    return Oo(t) || Oo(e)
  }
  function Bo(t) {
    if ('NHWC' === t) return 'channelsLast'
    if ('NCHW' === t) return 'channelsFirst'
    throw new Error('Unknown dataFormat ' + t)
  }
  function Po(t, e, n) {
    if ('complex64' === e) {
      if ('complex64' === t.dtype) return t.clone()
      var r = Gn(t.shape),
        o = t.toFloat(),
        a = n.complex(o, r)
      return r.dispose(), o.dispose(), a
    }
    if (!U(t.dtype, e)) return Lt.makeTensorFromDataId(t.dataId, t.shape, e)
    if ('complex64' === t.dtype) {
      var i = n.real(t)
      a = i.cast(e)
      return i.dispose(), a
    }
    if ('int32' === e) return n.int(t)
    if ('bool' === e) {
      var s = On(0, t.dtype)
      a = n.notEqual(t, s)
      return s.dispose(), a
    }
    throw new Error('Error in Cast: failed to cast ' + t.dtype + ' to ' + e)
  }
  function Lo(t, e) {
    return Lt.makeTensorFromDataId(t.dataId, e, t.dtype)
  }
  function Wo(t, e, n) {
    var r = (e - t) / (n - 1),
      o = tt(n, 'float32')
    o[0] = t
    for (var a = 1; a < o.length; a++) o[a] = o[a - 1] + r
    return Mn(o, 'float32')
  }
  var Uo = Object.freeze({
    castTensor: Po,
    reshapeTensor: Lo,
    linspaceImpl: Wo,
    upcastType: Dt,
    axesAreInnerMostDims: yn,
    combineLocations: xn,
    computeOutAndReduceShapes: bn,
    expandShapeToKeepDim: wn,
    assertAxesAreInnerMostDims: Cn,
    getAxesPermutation: En,
    getUndoAxesPermutation: Rn,
    getInnerMostAxes: In,
    getBroadcastDims: Co,
    getReductionAxes: Eo,
    assertAndGetBroadcastShape: Ro,
    assertParamsConsistent: kn,
    computeOutShape: Sn,
    computePool2DInfo: Io,
    computePool3DInfo: ko,
    computeConv2DInfo: So,
    computeConv3DInfo: Ao,
    computeDefaultPad: Do,
    tupleValuesAreOne: Oo,
    eitherStridesOrDilationsAreOne: Mo,
    convertConv2DDataFormat: Bo,
    PARALLELIZE_THRESHOLD: Yr,
    computeOptimalWindowSize: $r,
  })
  function Vo(t, e) {
    if (t.length !== e.length)
      throw new Error(
        'Cannot merge real and imag arrays of different lengths. real:' +
          t.length +
          ', imag: ' +
          e.length +
          '.'
      )
    for (var n = new Float32Array(2 * t.length), r = 0; r < n.length; r += 2)
      (n[r] = t[r / 2]), (n[r + 1] = e[r / 2])
    return n
  }
  function zo(t, e) {
    return { real: t[2 * e], imag: t[2 * e + 1] }
  }
  function Go(t, e, n, r) {
    ;(t[2 * r] = e), (t[2 * r + 1] = n)
  }
  function Ho(t, e, n) {
    var r = (n ? 2 : -2) * Math.PI * (t / e)
    return { real: Math.cos(r), imag: Math.sin(r) }
  }
  function qo(t, e, n) {
    var r = (function (t, e, n) {
        return (function (t, e, n) {
          var r = 0,
            o = t.length,
            a = 0,
            i = !1
          for (; r < o; ) {
            var s = n(e, t[(a = r + ((o - r) >>> 1))])
            s > 0 ? (r = a + 1) : ((o = a), (i = !s))
          }
          return i ? r : -r - 1
        })(t, e, n || Ko)
      })(t, e, n),
      o = r < 0 ? -(r + 1) : r
    t.splice(o, 0, e)
  }
  function Ko(t, e) {
    return t > e ? 1 : t < e ? -1 : 0
  }
  function jo(t, e, n, r, o) {
    return Yo(t, e, n, r, o, 0).selectedIndices
  }
  function Xo(t, e, n, r, o, a) {
    var i = Yo(t, e, n, r, o, a)
    return (
      i.numValidOutputs.dispose(),
      { selectedIndices: i.selectedIndices, selectedScores: i.selectedScores }
    )
  }
  function Yo(t, e, n, r, o, a, i, s) {
    void 0 === s && (s = !1)
    for (
      var u = Array.from(e)
          .map(function (t, e) {
            return { score: t, boxIndex: e, suppressBeginIndex: 0 }
          })
          .filter(function (t) {
            return t.score > o
          })
          .sort(Jo),
        c = a > 0 ? -0.5 / a : 0,
        l = [],
        h = [];
      l.length < n && u.length > 0;

    ) {
      var f = u.pop(),
        d = f.score,
        p = f.boxIndex,
        v = f.suppressBeginIndex
      if (d < o) break
      for (var m = !1, g = l.length - 1; g >= v; --g) {
        var y = $o(t, p, l[g])
        if (y >= r) {
          m = !0
          break
        }
        if (((f.score = f.score * Qo(r, c, y)), f.score <= o)) break
      }
      ;(f.suppressBeginIndex = l.length),
        m ||
          (f.score === d
            ? (l.push(p), h.push(f.score))
            : f.score > o && qo(u, f, Jo))
    }
    var x = l.length
    return (
      s && (l.fill(0, x), h.fill(0, x)),
      {
        selectedIndices: Mn(l, 'int32'),
        selectedScores: Mn(h, 'float32'),
        numValidOutputs: On(x, 'int32'),
      }
    )
  }
  function $o(t, e, n) {
    var r = t.subarray(4 * e, 4 * e + 4),
      o = t.subarray(4 * n, 4 * n + 4),
      a = Math.min(r[0], r[2]),
      i = Math.min(r[1], r[3]),
      s = Math.max(r[0], r[2]),
      u = Math.max(r[1], r[3]),
      c = Math.min(o[0], o[2]),
      l = Math.min(o[1], o[3]),
      h = Math.max(o[0], o[2]),
      f = Math.max(o[1], o[3]),
      d = (s - a) * (u - i),
      p = (h - c) * (f - l)
    if (d <= 0 || p <= 0) return 0
    var v = Math.max(a, c),
      m = Math.max(i, l),
      g = Math.min(s, h),
      y = Math.min(u, f),
      x = Math.max(g - v, 0) * Math.max(y - m, 0)
    return x / (d + p - x)
  }
  function Qo(t, e, n) {
    var r = Math.exp(e * n * n)
    return n <= t ? r : 0
  }
  function Jo(t, e) {
    return t.score - e.score || (t.score === e.score && e.boxIndex - t.boxIndex)
  }
  function Zo(t, e, n) {
    var r = new Array(t.rank).fill(0),
      o = t.shape.slice()
    return e.map(function (e) {
      o[n] = e
      var a = t.slice(r, o)
      return (r[n] += e), a
    })
  }
  function ta(t, e) {
    for (var n = new Array(t.rank), r = 0; r < n.length; r++)
      n[r] = t.shape[r] * e[r]
    var o = dr(n, t.dtype)
    for (r = 0; r < o.values.length; ++r) {
      for (
        var a = o.indexToLoc(r), i = new Array(t.rank), s = 0;
        s < i.length;
        s++
      )
        i[s] = a[s] % t.shape[s]
      var u = t.locToIndex(i)
      o.values[r] = t.values[u]
    }
    return o.toTensor()
  }
  function ea(t, e, n, r, o) {
    for (
      var a = e[e.length - 1],
        i = [t.length / a, a],
        s = i[0],
        u = i[1],
        c = B(n, s * r),
        l = B('int32', s * r),
        h = 0;
      h < s;
      h++
    ) {
      for (
        var f = h * u, d = t.subarray(f, f + u), p = [], v = 0;
        v < d.length;
        v++
      )
        p.push({ value: d[v], index: v })
      p.sort(function (t, e) {
        return e.value - t.value
      })
      var m = h * r,
        g = c.subarray(m, m + r),
        y = l.subarray(m, m + r)
      for (v = 0; v < r; v++) (g[v] = p[v].value), (y[v] = p[v].index)
    }
    var x = e.slice()
    return (x[x.length - 1] = r), [Fn(c, x, n), Fn(l, x, 'int32')]
  }
  function na(t, e) {
    for (var n = [], r = 0; r < e.length; r++) e[r] && n.push(r)
    var o = dr(t, 'int32'),
      a = dr([n.length, t.length], 'int32')
    for (r = 0; r < n.length; r++) {
      var i = o.indexToLoc(n[r]),
        s = r * t.length
      a.values.set(i, s)
    }
    return a.toTensor()
  }
  var ra = function (t, e) {
      ;(this.outputShape = []),
        (this.outputShape = t),
        (this.variableNames = e.map(function (t, e) {
          return 'T' + e
        }))
      var n = []
      this.variableNames.forEach(function (t) {
        n.push('float v' + t + ' = get' + t + 'AtOutCoords();')
      })
      var r = this.variableNames
        .map(function (t) {
          return 'v' + t
        })
        .join(' + ')
      this.userCode =
        '\n      void main() {\n        ' +
        n.join('\n        ') +
        '\n\n        float result = ' +
        r +
        ';\n        setOutput(result);\n      }\n    '
    },
    oa = function (t, e) {
      ;(this.outputShape = []),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = t),
        (this.variableNames = e.map(function (t, e) {
          return 'T' + e
        }))
      var n = []
      this.variableNames.forEach(function (t) {
        n.push('vec4 v' + t + ' = get' + t + 'AtOutCoords();')
      })
      var r = this.variableNames
        .map(function (t) {
          return 'v' + t
        })
        .join(' + ')
      this.userCode =
        '\n      void main() {\n        ' +
        n.join('\n        ') +
        '\n\n        vec4 result = ' +
        r +
        ';\n        setOutput(result);\n      }\n    '
    },
    aa = function (t, e, n) {
      this.variableNames = ['A']
      var r = t.windowSize,
        o = t.batchSize,
        a = t.inSize,
        i = Math.ceil(a / r)
      n || this.variableNames.push('bestIndicesA'), (this.outputShape = [o, i])
      var s = 'max' === e ? '>' : '<',
        u = n ? 'inOffset + i;' : 'round(getBestIndicesA(batch, inOffset + i));'
      this.userCode =
        '\n      void main() {\n        ivec2 coords = getOutputCoords();\n        int batch = coords[0];\n        int outIdx = coords[1];\n        int inOffset = outIdx * ' +
        r +
        ';\n\n        int bestIndex = inOffset;\n        float bestValue = getA(batch, bestIndex);\n\n        for (int i = 0; i < ' +
        r +
        '; i++) {\n          int inIdx = ' +
        u +
        ';\n          float candidate = getA(batch, inIdx);\n          if (candidate ' +
        s +
        ' bestValue) {\n            bestValue = candidate;\n            bestIndex = inIdx;\n          }\n        }\n        setOutput(float(bestIndex));\n      }\n    '
    }
  function ia(t, e) {
    return ['x', 'y', 'z', 'w', 'u', 'v'].slice(0, e).map(function (e) {
      return t + '.' + e
    })
  }
  function sa(t, e) {
    return 1 === e ? [t] : ia(t, e)
  }
  function ua() {
    var t, e, n, r, o, a, s, u, c, l
    return (
      2 === i().getNumber('WEBGL_VERSION')
        ? ((t = '#version 300 es'),
          (e = 'in'),
          (n = 'out'),
          (r = 'in'),
          (o = 'texture'),
          (a = 'outputColor'),
          (s = 'out vec4 outputColor;'),
          (u =
            '\n      bool isnan_custom(float val) {\n        return (val > 0.0 || val < 0.0) ? false : val != 0.0;\n      }\n\n      bvec4 isnan_custom(vec4 val) {\n        return bvec4(isnan_custom(val.x),\n          isnan_custom(val.y), isnan_custom(val.z), isnan_custom(val.w));\n      }\n\n      #define isnan(value) isnan_custom(value)\n    '),
          (c = ''),
          (l =
            '\n      #define round(value) newRound(value)\n      int newRound(float value) {\n        return int(floor(value + 0.5));\n      }\n\n      ivec4 newRound(vec4 value) {\n        return ivec4(floor(value + vec4(0.5)));\n      }\n    '))
        : ((t = ''),
          (e = 'attribute'),
          (n = 'varying'),
          (r = 'varying'),
          (o = 'texture2D'),
          (a = 'gl_FragColor'),
          (s = ''),
          (u =
            '\n      #define isnan(value) isnan_custom(value)\n      bool isnan_custom(float val) {\n        return (val > 0. || val < 1. || val == 0.) ? false : true;\n      }\n      bvec4 isnan_custom(vec4 val) {\n        return bvec4(isnan(val.x), isnan(val.y), isnan(val.z), isnan(val.w));\n      }\n    '),
          (c =
            '\n      uniform float INFINITY;\n\n      bool isinf(float val) {\n        return abs(val) == INFINITY;\n      }\n      bvec4 isinf(vec4 val) {\n        return equal(abs(val), vec4(INFINITY));\n      }\n    '),
          (l =
            '\n      int round(float value) {\n        return int(floor(value + 0.5));\n      }\n\n      ivec4 round(vec4 value) {\n        return ivec4(floor(value + vec4(0.5)));\n      }\n    ')),
      {
        version: t,
        attribute: e,
        varyingVs: n,
        varyingFs: r,
        texture2D: o,
        output: a,
        defineOutput: s,
        defineSpecialNaN: u,
        defineSpecialInf: c,
        defineRound: l,
      }
    )
  }
  function ca(t, e, n) {
    void 0 === n && (n = 'index')
    var r = $(e)
    return r
      .map(function (e, o) {
        return (
          'int ' +
          t[o] +
          ' = ' +
          n +
          ' / ' +
          e +
          '; ' +
          (o === r.length - 1
            ? 'int ' + t[o + 1] + ' = ' + n + ' - ' + t[o] + ' * ' + e
            : 'index -= ' + t[o] + ' * ' + e) +
          ';'
        )
      })
      .join('')
  }
  function la(t) {
    var e = $(t).map(function (t) {
      return t.toString()
    })
    return (
      '\n  int getFlatIndex(ivec3 coords) {\n    return coords.x * ' +
      e[0] +
      ' + coords.y * ' +
      e[1] +
      ' + coords.z;\n  }\n'
    )
  }
  var ha =
    '\n  const float FLOAT_MAX = 1.70141184e38;\n  const float FLOAT_MIN = 1.17549435e-38;\n\n  lowp vec4 encode_float(highp float v) {\n    if (isnan(v)) {\n      return vec4(255, 255, 255, 255);\n    }\n\n    highp float av = abs(v);\n\n    if(av < FLOAT_MIN) {\n      return vec4(0.0, 0.0, 0.0, 0.0);\n    } else if(v > FLOAT_MAX) {\n      return vec4(0.0, 0.0, 128.0, 127.0) / 255.0;\n    } else if(v < -FLOAT_MAX) {\n      return vec4(0.0, 0.0,  128.0, 255.0) / 255.0;\n    }\n\n    highp vec4 c = vec4(0,0,0,0);\n\n    highp float e = floor(log2(av));\n    highp float m = exp2(fract(log2(av))) - 1.0;\n\n    c[2] = floor(128.0 * m);\n    m -= c[2] / 128.0;\n    c[1] = floor(32768.0 * m);\n    m -= c[1] / 32768.0;\n    c[0] = floor(8388608.0 * m);\n\n    highp float ebias = e + 127.0;\n    c[3] = floor(ebias / 2.0);\n    ebias -= c[3] * 2.0;\n    c[2] += floor(ebias) * 128.0;\n\n    c[3] += 128.0 * step(0.0, -v);\n\n    return c / 255.0;\n  }\n'
  function fa(t, e, n, r) {
    var o = []
    t.forEach(function (t) {
      var e = k(t.shapeInfo.logicalShape)
      t.shapeInfo.isUniform
        ? o.push('uniform float ' + t.name + (e > 1 ? '[' + e + ']' : '') + ';')
        : (o.push('uniform sampler2D ' + t.name + ';'),
          o.push('uniform int offset' + t.name + ';'))
    })
    var a,
      i,
      s = o.join('\n'),
      u = t
        .map(function (t) {
          return (function (t, e, n) {
            void 0 === n && (n = !1)
            var r = ''
            r += n ? pa(t) : da(t)
            var o = t.shapeInfo.logicalShape,
              a = e.logicalShape
            o.length <= a.length &&
              (r += n
                ? (function (t, e) {
                    var n,
                      r = t.name,
                      o = r.charAt(0).toUpperCase() + r.slice(1),
                      a = 'get' + o + 'AtOutCoords',
                      i = t.shapeInfo.logicalShape.length,
                      s = e.logicalShape.length,
                      u = Co(t.shapeInfo.logicalShape, e.logicalShape),
                      c = wa(s),
                      l = s - i,
                      h = ['x', 'y', 'z', 'w', 'u', 'v']
                    n =
                      0 === i
                        ? ''
                        : s < 2 && u.length >= 1
                        ? 'coords = 0;'
                        : u
                            .map(function (t) {
                              return 'coords.' + h[t + l] + ' = 0;'
                            })
                            .join('\n')
                    var f = ''
                    f =
                      s < 2 && i > 0
                        ? 'coords'
                        : t.shapeInfo.logicalShape
                            .map(function (t, e) {
                              return 'coords.' + h[e + l]
                            })
                            .join(', ')
                    var d = 'return outputValue;',
                      p = 1 === k(t.shapeInfo.logicalShape),
                      v = 1 === k(e.logicalShape)
                    if (1 !== i || p || v) {
                      if (p && !v)
                        d =
                          1 === s
                            ? '\n        return vec4(outputValue.x, outputValue.x, 0., 0.);\n      '
                            : '\n        return vec4(outputValue.x);\n      '
                      else if (u.length) {
                        var m = i - 2,
                          g = i - 1
                        u.indexOf(m) > -1 && u.indexOf(g) > -1
                          ? (d = 'return vec4(outputValue.x);')
                          : u.indexOf(m) > -1
                          ? (d =
                              'return vec4(outputValue.x, outputValue.y, outputValue.x, outputValue.y);')
                          : u.indexOf(g) > -1 &&
                            (d = 'return vec4(outputValue.xx, outputValue.zz);')
                      }
                    } else
                      d =
                        '\n      return vec4(outputValue.xy, outputValue.xy);\n    '
                    return (
                      '\n    vec4 ' +
                      a +
                      '() {\n      ' +
                      c +
                      ' coords = getOutputCoords();\n      ' +
                      n +
                      '\n      vec4 outputValue = get' +
                      o +
                      '(' +
                      f +
                      ');\n      ' +
                      d +
                      '\n    }\n  '
                    )
                  })(t, e)
                : (function (t, e) {
                    var n = t.name,
                      r = n.charAt(0).toUpperCase() + n.slice(1),
                      o = 'get' + r + 'AtOutCoords',
                      a = e.texShape,
                      i = t.shapeInfo.texShape,
                      s = t.shapeInfo.logicalShape.length,
                      u = e.logicalShape.length
                    if (
                      !t.shapeInfo.isUniform &&
                      s === u &&
                      null == t.shapeInfo.flatOffset &&
                      S(i, a)
                    )
                      return (
                        '\n      float ' +
                        o +
                        '() {\n        return sampleTexture(' +
                        n +
                        ', resultUV);\n      }\n    '
                      )
                    var c,
                      l = wa(u),
                      h = Co(t.shapeInfo.logicalShape, e.logicalShape),
                      f = u - s,
                      d = ['x', 'y', 'z', 'w', 'u', 'v']
                    c =
                      0 === s
                        ? ''
                        : u < 2 && h.length >= 1
                        ? 'coords = 0;'
                        : h
                            .map(function (t) {
                              return 'coords.' + d[t + f] + ' = 0;'
                            })
                            .join('\n')
                    var p = ''
                    p =
                      u < 2 && s > 0
                        ? 'coords'
                        : t.shapeInfo.logicalShape
                            .map(function (t, e) {
                              return 'coords.' + d[e + f]
                            })
                            .join(', ')
                    return (
                      '\n    float ' +
                      o +
                      '() {\n      ' +
                      l +
                      ' coords = getOutputCoords();\n      ' +
                      c +
                      '\n      return get' +
                      r +
                      '(' +
                      p +
                      ');\n    }\n  '
                    )
                  })(t, e))
            return r
          })(t, e, r)
        })
        .join('\n'),
      c = e.texShape,
      l = ua(),
      h = (function (t) {
        return (
          '\n    float sampleTexture(sampler2D textureSampler, vec2 uv) {\n      return ' +
          t.texture2D +
          '(textureSampler, uv).r;\n    }\n  '
        )
      })(l),
      f = (function (t) {
        return (
          t.version +
          '\n    precision highp float;\n    precision highp int;\n    precision highp sampler2D;\n    ' +
          t.varyingFs +
          ' vec2 resultUV;\n    ' +
          t.defineOutput +
          '\n    const vec2 halfCR = vec2(0.5, 0.5);\n\n    struct ivec5\n    {\n      int x;\n      int y;\n      int z;\n      int w;\n      int u;\n    };\n\n    struct ivec6\n    {\n      int x;\n      int y;\n      int z;\n      int w;\n      int u;\n      int v;\n    };\n\n    uniform float NAN;\n    ' +
          t.defineSpecialNaN +
          '\n    ' +
          t.defineSpecialInf +
          '\n    ' +
          t.defineRound +
          '\n\n    int imod(int x, int y) {\n      return x - y * (x / y);\n    }\n\n    int idiv(int a, int b, float sign) {\n      int res = a / b;\n      int mod = imod(a, b);\n      if (sign < 0. && mod != 0) {\n        res -= 1;\n      }\n      return res;\n    }\n\n    //Based on the work of Dave Hoskins\n    //https://www.shadertoy.com/view/4djSRW\n    #define HASHSCALE1 443.8975\n    float random(float seed){\n      vec2 p = resultUV * seed;\n      vec3 p3  = fract(vec3(p.xyx) * HASHSCALE1);\n      p3 += dot(p3, p3.yzx + 19.19);\n      return fract((p3.x + p3.y) * p3.z);\n    }\n\n    ' +
          va +
          '\n    ' +
          ma +
          '\n    ' +
          ga +
          '\n  '
        )
      })(l)
    return (
      e.isPacked
        ? ((a = (function (t, e) {
            switch (t.length) {
              case 0:
                return '\n    int getOutputCoords() {\n      return 0;\n    }\n  '
              case 1:
                return (function (t, e) {
                  var n = [Math.ceil(e[0] / 2), Math.ceil(e[1] / 2)]
                  if (1 === n[0])
                    return (
                      '\n      int getOutputCoords() {\n        return 2 * int(resultUV.x * ' +
                      n[1] +
                      '.0);\n      }\n    '
                    )
                  if (1 === n[1])
                    return (
                      '\n      int getOutputCoords() {\n        return 2 * int(resultUV.y * ' +
                      n[0] +
                      '.0);\n      }\n    '
                    )
                  return (
                    '\n    int getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    n[0] +
                    ', ' +
                    n[1] +
                    '));\n      return 2 * (resTexRC.x * ' +
                    n[1] +
                    ' + resTexRC.y);\n    }\n  '
                  )
                })(0, e)
              case 2:
                return (function (t, e) {
                  var n = [Math.ceil(e[0] / 2), Math.ceil(e[1] / 2)]
                  if (S(t, e))
                    return (
                      '\n      ivec2 getOutputCoords() {\n        return 2 * ivec2(resultUV.yx * vec2(' +
                      n[0] +
                      ', ' +
                      n[1] +
                      '));\n      }\n    '
                    )
                  var r = Math.ceil(t[1] / 2)
                  return (
                    '\n    ivec2 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    n[0] +
                    ', ' +
                    n[1] +
                    '));\n\n      int index = resTexRC.x * ' +
                    n[1] +
                    ' + resTexRC.y;\n      int r = 2 * (index / ' +
                    r +
                    ');\n      int c = imod(index, ' +
                    r +
                    ') * 2;\n\n      return ivec2(r, c);\n    }\n  '
                  )
                })(t, e)
              case 3:
                return (
                  (n = t),
                  (r = e),
                  (o = [Math.ceil(r[0] / 2), Math.ceil(r[1] / 2)]),
                  (a = Math.ceil(n[2] / 2)),
                  (i = a * Math.ceil(n[1] / 2)),
                  '\n    ivec3 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    o[0] +
                    ', ' +
                    o[1] +
                    '));\n      int index = resTexRC.x * ' +
                    o[1] +
                    ' + resTexRC.y;\n\n      int b = index / ' +
                    i +
                    ';\n      index -= b * ' +
                    i +
                    ';\n\n      int r = 2 * (index / ' +
                    a +
                    ');\n      int c = imod(index, ' +
                    a +
                    ') * 2;\n\n      return ivec3(b, r, c);\n    }\n  '
                )
              default:
                return (function (t, e) {
                  for (
                    var n = [Math.ceil(e[0] / 2), Math.ceil(e[1] / 2)],
                      r = Math.ceil(t[t.length - 1] / 2),
                      o = r * Math.ceil(t[t.length - 2] / 2),
                      a = o,
                      i = '',
                      s = 'b, r, c',
                      u = 2;
                    u < t.length - 1;
                    u++
                  )
                    (a *= t[t.length - u - 1]),
                      (i =
                        '\n      int b' +
                        u +
                        ' = index / ' +
                        a +
                        ';\n      index -= b' +
                        u +
                        ' * ' +
                        a +
                        ';\n    ' +
                        i),
                      (s = 'b' + u + ', ' + s)
                  return (
                    '\n    ivec' +
                    t.length +
                    ' getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    n[0] +
                    ', ' +
                    n[1] +
                    '));\n      int index = resTexRC.x * ' +
                    n[1] +
                    ' + resTexRC.y;\n\n      ' +
                    i +
                    '\n\n      int b = index / ' +
                    o +
                    ';\n      index -= b * ' +
                    o +
                    ';\n\n      int r = 2 * (index / ' +
                    r +
                    ');\n      int c = imod(index, ' +
                    r +
                    ') * 2;\n\n      return ivec' +
                    t.length +
                    '(' +
                    s +
                    ');\n    }\n  '
                  )
                })(t, e)
            }
            var n, r, o, a, i
          })(e.logicalShape, c)),
          (i = (function (t) {
            return (
              '\n    void setOutput(vec4 val) {\n      ' +
              t.output +
              ' = val;\n    }\n  '
            )
          })(l)))
        : ((a = (function (t, e) {
            switch (t.length) {
              case 0:
                return '\n    int getOutputCoords() {\n      return 0;\n    }\n  '
              case 1:
                return (function (t, e) {
                  if (1 === e[0])
                    return (
                      '\n      int getOutputCoords() {\n        return int(resultUV.x * ' +
                      e[1] +
                      '.0);\n      }\n    '
                    )
                  if (1 === e[1])
                    return (
                      '\n      int getOutputCoords() {\n        return int(resultUV.y * ' +
                      e[0] +
                      '.0);\n      }\n    '
                    )
                  return (
                    '\n    int getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    e[0] +
                    ', ' +
                    e[1] +
                    '));\n      return resTexRC.x * ' +
                    e[1] +
                    ' + resTexRC.y;\n    }\n  '
                  )
                })(0, e)
              case 2:
                return (function (t, e) {
                  if (S(t, e))
                    return (
                      '\n      ivec2 getOutputCoords() {\n        return ivec2(resultUV.yx * vec2(' +
                      e[0] +
                      ', ' +
                      e[1] +
                      '));\n      }\n    '
                    )
                  if (1 === t[1])
                    return (
                      '\n      ivec2 getOutputCoords() {\n        ivec2 resTexRC = ivec2(resultUV.yx *\n                               vec2(' +
                      e[0] +
                      ', ' +
                      e[1] +
                      '));\n        int index = resTexRC.x * ' +
                      e[1] +
                      ' + resTexRC.y;\n        return ivec2(index, 0);\n      }\n    '
                    )
                  if (1 === t[0])
                    return (
                      '\n      ivec2 getOutputCoords() {\n        ivec2 resTexRC = ivec2(resultUV.yx *\n                               vec2(' +
                      e[0] +
                      ', ' +
                      e[1] +
                      '));\n        int index = resTexRC.x * ' +
                      e[1] +
                      ' + resTexRC.y;\n        return ivec2(0, index);\n      }\n    '
                    )
                  return (
                    '\n    ivec2 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    e[0] +
                    ', ' +
                    e[1] +
                    '));\n      int index = resTexRC.x * ' +
                    e[1] +
                    ' + resTexRC.y;\n      int r = index / ' +
                    t[1] +
                    ';\n      int c = index - r * ' +
                    t[1] +
                    ';\n      return ivec2(r, c);\n    }\n  '
                  )
                })(t, e)
              case 3:
                return (
                  (n = e),
                  (r = ca(['r', 'c', 'd'], t)),
                  '\n    ivec3 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n                             vec2(' +
                    n[0] +
                    ', ' +
                    n[1] +
                    '));\n      int index = resTexRC.x * ' +
                    n[1] +
                    ' + resTexRC.y;\n      ' +
                    r +
                    '\n      return ivec3(r, c, d);\n    }\n  '
                )
              case 4:
                return (function (t, e) {
                  var n = ca(['r', 'c', 'd', 'd2'], t)
                  return (
                    '\n    ivec4 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n        vec2(' +
                    e[0] +
                    ', ' +
                    e[1] +
                    '));\n      int index = resTexRC.x * ' +
                    e[1] +
                    ' + resTexRC.y;\n      ' +
                    n +
                    '\n      return ivec4(r, c, d, d2);\n    }\n  '
                  )
                })(t, e)
              case 5:
                return (function (t, e) {
                  var n = ca(['r', 'c', 'd', 'd2', 'd3'], t)
                  return (
                    '\n    ivec5 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx * vec2(' +
                    e[0] +
                    ',\n                             ' +
                    e[1] +
                    '));\n\n      int index = resTexRC.x * ' +
                    e[1] +
                    ' + resTexRC.y;\n\n      ' +
                    n +
                    '\n\n      ivec5 outShape = ivec5(r, c, d, d2, d3);\n      return outShape;\n    }\n  '
                  )
                })(t, e)
              case 6:
                return (function (t, e) {
                  var n = ca(['r', 'c', 'd', 'd2', 'd3', 'd4'], t)
                  return (
                    '\n    ivec6 getOutputCoords() {\n      ivec2 resTexRC = ivec2(resultUV.yx *\n        vec2(' +
                    e[0] +
                    ', ' +
                    e[1] +
                    '));\n      int index = resTexRC.x * ' +
                    e[1] +
                    ' + resTexRC.y;\n\n      ' +
                    n +
                    '\n\n      ivec6 result = ivec6(r, c, d, d2, d3, d4);\n      return result;\n    }\n  '
                  )
                })(t, e)
              default:
                throw new Error(
                  t.length + '-D output sampling is not yet supported'
                )
            }
            var n, r
          })(e.logicalShape, c)),
          (i = (function (t) {
            return (
              '\n    void setOutput(float val) {\n      ' +
              t.output +
              ' = vec4(val, 0, 0, 0);\n    }\n  '
            )
          })(l))),
      r && (f += ya),
      [f, h, i, s, a, u, n].join('\n')
    )
  }
  function da(t) {
    var e = t.shapeInfo.logicalShape
    switch (e.length) {
      case 0:
        return (function (t) {
          var e = t.name,
            n = 'get' + e.charAt(0).toUpperCase() + e.slice(1)
          if (t.shapeInfo.isUniform)
            return 'float ' + n + '() {return ' + e + ';}'
          var r = t.shapeInfo.texShape,
            o = r[0],
            a = r[1]
          if (1 === o && 1 === a)
            return (
              '\n      float ' +
              n +
              '() {\n        return sampleTexture(' +
              e +
              ', halfCR);\n      }\n    '
            )
          var i = t.shapeInfo.texShape,
            s = i[0],
            u = i[1],
            c = xa(e)
          return (
            '\n    float ' +
            n +
            '() {\n      vec2 uv = uvFromFlat(' +
            s +
            ', ' +
            u +
            ', ' +
            c +
            ');\n      return sampleTexture(' +
            e +
            ', uv);\n    }\n  '
          )
        })(t)
      case 1:
        return (function (t) {
          var e = t.name,
            n = 'get' + e.charAt(0).toUpperCase() + e.slice(1)
          if (t.shapeInfo.isUniform)
            return (
              '\n      float ' +
              n +
              '(int index) {\n        ' +
              ba(t) +
              '\n      }\n    '
            )
          var r = t.shapeInfo.texShape,
            o = r[0],
            a = r[1]
          if (1 === a && 1 === o)
            return (
              '\n      float ' +
              n +
              '(int index) {\n        return sampleTexture(' +
              e +
              ', halfCR);\n      }\n    '
            )
          var i = xa(e)
          if (1 === a)
            return (
              '\n      float ' +
              n +
              '(int index) {\n        vec2 uv = vec2(0.5, (float(index + ' +
              i +
              ') + 0.5) / ' +
              o +
              '.0);\n        return sampleTexture(' +
              e +
              ', uv);\n      }\n    '
            )
          if (1 === o)
            return (
              '\n      float ' +
              n +
              '(int index) {\n        vec2 uv = vec2((float(index + ' +
              i +
              ') + 0.5) / ' +
              a +
              '.0, 0.5);\n        return sampleTexture(' +
              e +
              ', uv);\n      }\n    '
            )
          return (
            '\n    float ' +
            n +
            '(int index) {\n      vec2 uv = uvFromFlat(' +
            o +
            ', ' +
            a +
            ', index + ' +
            i +
            ');\n      return sampleTexture(' +
            e +
            ', uv);\n    }\n  '
          )
        })(t)
      case 2:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = t.shapeInfo.texShape
          if (null != o && S(e, o)) {
            var a = o[0],
              i = o[1]
            return (
              '\n    float ' +
              r +
              '(int row, int col) {\n      vec2 uv = (vec2(col, row) + halfCR) / vec2(' +
              i +
              '.0, ' +
              a +
              '.0);\n      return sampleTexture(' +
              n +
              ', uv);\n    }\n  '
            )
          }
          var s = M(e),
            u = s.newShape,
            c = s.keptDims,
            l = u
          if (l.length < e.length) {
            var h = Ca(t, l)
            return (
              '\n      ' +
              da(h) +
              '\n      float ' +
              r +
              '(int row, int col) {\n        return ' +
              r +
              '(' +
              Ea(['row', 'col'], c) +
              ');\n      }\n    '
            )
          }
          if (t.shapeInfo.isUniform)
            return (
              '\n      float ' +
              r +
              '(int row, int col) {\n        int index = round(dot(vec2(row, col), vec2(' +
              e[1] +
              ', 1)));\n        ' +
              ba(t) +
              '\n      }\n    '
            )
          var f = o[0],
            d = o[1],
            p = xa(n)
          if (1 === d)
            return (
              '\n    float ' +
              r +
              '(int row, int col) {\n      float index = dot(vec3(row, col, ' +
              p +
              '), vec3(' +
              e[1] +
              ', 1, 1));\n      vec2 uv = vec2(0.5, (index + 0.5) / ' +
              f +
              '.0);\n      return sampleTexture(' +
              n +
              ', uv);\n    }\n  '
            )
          if (1 === f)
            return (
              '\n    float ' +
              r +
              '(int row, int col) {\n      float index = dot(vec3(row, col, ' +
              p +
              '), vec3(' +
              e[1] +
              ', 1, 1));\n      vec2 uv = vec2((index + 0.5) / ' +
              d +
              '.0, 0.5);\n      return sampleTexture(' +
              n +
              ', uv);\n    }\n  '
            )
          return (
            '\n  float ' +
            r +
            '(int row, int col) {\n    // Explicitly use integer operations as dot() only works on floats.\n    int index = row * ' +
            e[1] +
            ' + col + ' +
            p +
            ';\n    vec2 uv = uvFromFlat(' +
            f +
            ', ' +
            d +
            ', index);\n    return sampleTexture(' +
            n +
            ', uv);\n  }\n'
          )
        })(t)
      case 3:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = e[1] * e[2],
            a = e[2],
            i = M(e),
            s = i.newShape,
            u = i.keptDims,
            c = s
          if (c.length < e.length) {
            var l = Ca(t, c)
            return (
              '\n        ' +
              da(l) +
              '\n        float ' +
              r +
              '(int row, int col, int depth) {\n          return ' +
              r +
              '(' +
              Ea(['row', 'col', 'depth'], u) +
              ');\n        }\n      '
            )
          }
          if (t.shapeInfo.isUniform)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth) {\n        int index = round(dot(vec3(row, col, depth),\n                          vec3(' +
              o +
              ', ' +
              a +
              ', 1)));\n        ' +
              ba(t) +
              '\n      }\n    '
            )
          var h = t.shapeInfo.texShape,
            f = h[0],
            d = h[1],
            p = t.shapeInfo.flatOffset
          if (d === o && null == p)
            return (
              '\n        float ' +
              r +
              '(int row, int col, int depth) {\n          float texR = float(row);\n          float texC = dot(vec2(col, depth), vec2(' +
              a +
              ', 1));\n          vec2 uv = (vec2(texC, texR) + halfCR) /\n                     vec2(' +
              d +
              '.0, ' +
              f +
              '.0);\n          return sampleTexture(' +
              n +
              ', uv);\n        }\n      '
            )
          if (d === a && null == p)
            return (
              '\n    float ' +
              r +
              '(int row, int col, int depth) {\n      float texR = dot(vec2(row, col), vec2(' +
              e[1] +
              ', 1));\n      float texC = float(depth);\n      vec2 uv = (vec2(texC, texR) + halfCR) / vec2(' +
              d +
              '.0, ' +
              f +
              '.0);\n      return sampleTexture(' +
              n +
              ', uv);\n    }\n  '
            )
          var v = xa(n)
          return (
            '\n      float ' +
            r +
            '(int row, int col, int depth) {\n        // Explicitly use integer operations as dot() only works on floats.\n        int index = row * ' +
            o +
            ' + col * ' +
            a +
            ' + depth + ' +
            v +
            ';\n        vec2 uv = uvFromFlat(' +
            f +
            ', ' +
            d +
            ', index);\n        return sampleTexture(' +
            n +
            ', uv);\n      }\n  '
          )
        })(t)
      case 4:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = e[3],
            a = e[2] * o,
            i = e[1] * a,
            s = M(e),
            u = s.newShape,
            c = s.keptDims
          if (u.length < e.length) {
            var l = Ca(t, u)
            return (
              '\n      ' +
              da(l) +
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2) {\n        return ' +
              r +
              '(' +
              Ea(['row', 'col', 'depth', 'depth2'], c) +
              ');\n      }\n    '
            )
          }
          if (t.shapeInfo.isUniform)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2) {\n        int index = round(dot(vec4(row, col, depth, depth2),\n                          vec4(' +
              i +
              ', ' +
              a +
              ', ' +
              o +
              ', 1)));\n        ' +
              ba(t) +
              '\n      }\n    '
            )
          var h = t.shapeInfo.flatOffset,
            f = t.shapeInfo.texShape,
            d = f[0],
            p = f[1]
          if (p === i && null == h)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2) {\n        float texR = float(row);\n        float texC =\n            dot(vec3(col, depth, depth2),\n                vec3(' +
              a +
              ', ' +
              o +
              ', 1));\n        vec2 uv = (vec2(texC, texR) + halfCR) /\n                   vec2(' +
              p +
              '.0, ' +
              d +
              '.0);\n        return sampleTexture(' +
              n +
              ', uv);\n      }\n    '
            )
          if (p === o && null == h)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2) {\n        float texR = dot(vec3(row, col, depth),\n                         vec3(' +
              e[1] * e[2] +
              ', ' +
              e[2] +
              ', 1));\n        float texC = float(depth2);\n        vec2 uv = (vec2(texC, texR) + halfCR) /\n                  vec2(' +
              p +
              '.0, ' +
              d +
              '.0);\n        return sampleTexture(' +
              n +
              ', uv);\n      }\n    '
            )
          var v = xa(n)
          return (
            '\n    float ' +
            r +
            '(int row, int col, int depth, int depth2) {\n      // Explicitly use integer operations as dot() only works on floats.\n      int index = row * ' +
            i +
            ' + col * ' +
            a +
            ' +\n          depth * ' +
            o +
            ' + depth2;\n      vec2 uv = uvFromFlat(' +
            d +
            ', ' +
            p +
            ', index + ' +
            v +
            ');\n      return sampleTexture(' +
            n +
            ', uv);\n    }\n  '
          )
        })(t)
      case 5:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = e[4],
            a = e[3] * o,
            i = e[2] * a,
            s = e[1] * i,
            u = M(e),
            c = u.newShape,
            l = u.keptDims
          if (c.length < e.length) {
            var h = Ca(t, c)
            return (
              '\n      ' +
              da(h) +
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2, int depth3) {\n        return ' +
              r +
              '(' +
              Ea(['row', 'col', 'depth', 'depth2', 'depth3'], l) +
              ');\n      }\n    '
            )
          }
          if (t.shapeInfo.isUniform)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2, int depth3) {\n        float index = dot(\n          vec4(row, col, depth, depth2),\n          vec4(' +
              s +
              ', ' +
              i +
              ', ' +
              a +
              ', ' +
              o +
              ')) +\n          depth3;\n        ' +
              ba(t) +
              '\n      }\n    '
            )
          var f = t.shapeInfo.flatOffset,
            d = t.shapeInfo.texShape,
            p = d[0],
            v = d[1]
          if (v === s && null == f)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2, int depth3) {\n        int texR = row;\n        float texC = dot(vec4(col, depth, depth2, depth3),\n                         vec4(' +
              i +
              ', ' +
              a +
              ', ' +
              o +
              ', 1));\n        vec2 uv = (vec2(texC, texR) + halfCR) /\n                   vec2(' +
              v +
              '.0, ' +
              p +
              '.0);\n        return sampleTexture(' +
              n +
              ', uv);\n      }\n    '
            )
          if (v === o && null == f)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth, int depth2, int depth3) {\n        float texR = dot(\n          vec4(row, col, depth, depth2),\n          vec4(' +
              e[1] * e[2] * e[3] +
              ',\n               ' +
              e[2] * e[3] +
              ', ' +
              e[3] +
              ', 1));\n        int texC = depth3;\n        vec2 uv = (vec2(texC, texR) + halfCR) /\n                  vec2(' +
              v +
              '.0, ' +
              p +
              '.0);\n        return sampleTexture(' +
              n +
              ', uv);\n      }\n    '
            )
          var m = xa(n)
          return (
            '\n    float ' +
            r +
            '(int row, int col, int depth, int depth2, int depth3) {\n      // Explicitly use integer operations as dot() only works on floats.\n      int index = row * ' +
            s +
            ' + col * ' +
            i +
            ' + depth * ' +
            a +
            ' +\n          depth2 * ' +
            o +
            ' + depth3 + ' +
            m +
            ';\n      vec2 uv = uvFromFlat(' +
            p +
            ', ' +
            v +
            ', index);\n      return sampleTexture(' +
            n +
            ', uv);\n    }\n  '
          )
        })(t)
      case 6:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = M(e),
            a = o.newShape,
            i = o.keptDims
          if (a.length < e.length) {
            var s = Ca(t, a)
            return (
              '\n      ' +
              da(s) +
              '\n      float ' +
              r +
              '(int row, int col, int depth,\n                    int depth2, int depth3, int depth4) {\n        return ' +
              r +
              '(' +
              Ea(['row', 'col', 'depth', 'depth2', 'depth3', 'depth4'], i) +
              ');\n      }\n    '
            )
          }
          var u = e[5],
            c = e[4] * u,
            l = e[3] * c,
            h = e[2] * l,
            f = e[1] * h
          if (t.shapeInfo.isUniform)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth,\n                  int depth2, int depth3, int depth4) {\n        int index = round(dot(\n          vec4(row, col, depth, depth2),\n          vec4(' +
              f +
              ', ' +
              h +
              ', ' +
              l +
              ', ' +
              c +
              ')) +\n          dot(\n            vec2(depth3, depth4),\n            vec2(' +
              u +
              ', 1)));\n        ' +
              ba(t) +
              '\n      }\n    '
            )
          var d = t.shapeInfo.flatOffset,
            p = t.shapeInfo.texShape,
            v = p[0],
            m = p[1]
          if (m === f && null == d)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth,\n                    int depth2, int depth3, int depth4) {\n        int texR = row;\n        float texC = dot(vec4(col, depth, depth2, depth3),\n          vec4(' +
              h +
              ', ' +
              l +
              ', ' +
              c +
              ', ' +
              u +
              ')) +\n               float(depth4);\n        vec2 uv = (vec2(texC, texR) + halfCR) /\n                   vec2(' +
              m +
              '.0, ' +
              v +
              '.0);\n        return sampleTexture(' +
              n +
              ', uv);\n      }\n    '
            )
          if (m === u && null == d)
            return (
              '\n      float ' +
              r +
              '(int row, int col, int depth,\n                    int depth2, int depth3, int depth4) {\n        float texR = dot(vec4(row, col, depth, depth2),\n          vec4(' +
              e[1] * e[2] * e[3] * e[4] +
              ',\n               ' +
              e[2] * e[3] * e[4] +
              ',\n               ' +
              e[3] * e[4] +
              ',\n               ' +
              e[4] +
              ')) + float(depth3);\n        int texC = depth4;\n        vec2 uv = (vec2(texC, texR) + halfCR) /\n                  vec2(' +
              m +
              '.0, ' +
              v +
              '.0);\n        return sampleTexture(' +
              n +
              ', uv);\n      }\n    '
            )
          var g = xa(n)
          return (
            '\n    float ' +
            r +
            '(int row, int col, int depth,\n                  int depth2, int depth3, int depth4) {\n      // Explicitly use integer operations as dot() only works on floats.\n      int index = row * ' +
            f +
            ' + col * ' +
            h +
            ' + depth * ' +
            l +
            ' +\n          depth2 * ' +
            c +
            ' + depth3 * ' +
            u +
            ' + depth4 + ' +
            g +
            ';\n      vec2 uv = uvFromFlat(' +
            v +
            ', ' +
            m +
            ', index);\n      return sampleTexture(' +
            n +
            ', uv);\n    }\n  '
          )
        })(t)
      default:
        throw new Error(e.length + '-D input sampling is not yet supported')
    }
  }
  function pa(t) {
    var e, n, r
    switch (t.shapeInfo.logicalShape.length) {
      case 0:
        return (
          (e = t.name),
          (n = 'get' + e.charAt(0).toUpperCase() + e.slice(1)),
          (r = ua()),
          '\n    vec4 ' +
            n +
            '() {\n      return ' +
            r.texture2D +
            '(' +
            e +
            ', halfCR);\n    }\n  '
        )
      case 1:
        return (function (t) {
          var e = t.name,
            n = 'get' + e.charAt(0).toUpperCase() + e.slice(1),
            r = t.shapeInfo.texShape,
            o = [Math.ceil(r[0] / 2), Math.ceil(r[1] / 2)],
            a = ua()
          return (
            '\n    vec4 ' +
            n +
            '(int index) {\n      vec2 uv = packedUVfrom1D(\n        ' +
            o[0] +
            ', ' +
            o[1] +
            ', index);\n      return ' +
            a.texture2D +
            '(' +
            e +
            ', uv);\n    }\n  '
          )
        })(t)
      case 2:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = t.shapeInfo.texShape,
            a = o[0],
            i = o[1],
            s = ua()
          if (null != o && S(e, o))
            return (
              '\n      vec4 ' +
              r +
              '(int row, int col) {\n        vec2 uv = (vec2(col, row) + halfCR) / vec2(' +
              i +
              '.0, ' +
              a +
              '.0);\n\n        return ' +
              s.texture2D +
              '(' +
              n +
              ', uv);\n      }\n    '
            )
          var u = [Math.ceil(o[0] / 2), Math.ceil(o[1] / 2)],
            c = Math.ceil(e[1] / 2)
          return (
            '\n    vec4 ' +
            r +
            '(int row, int col) {\n      vec2 uv = packedUVfrom2D(' +
            c +
            ', ' +
            u[0] +
            ', ' +
            u[1] +
            ', row, col);\n      return ' +
            s.texture2D +
            '(' +
            n +
            ', uv);\n    }\n  '
          )
        })(t)
      case 3:
        return (function (t) {
          var e = t.shapeInfo.logicalShape,
            n = t.name,
            r = 'get' + n.charAt(0).toUpperCase() + n.slice(1),
            o = t.shapeInfo.texShape,
            a = [Math.ceil(o[0] / 2), Math.ceil(o[1] / 2)]
          if (1 === e[0]) {
            var i = e.slice(1),
              s = Ca(t, i)
            return (
              '\n        ' +
              pa(s) +
              '\n        vec4 ' +
              r +
              '(int b, int row, int col) {\n          return ' +
              r +
              '(' +
              Ea(['b', 'row', 'col'], [1, 2]) +
              ');\n        }\n      '
            )
          }
          var u = a[0],
            c = a[1],
            l = Math.ceil(e[2] / 2),
            h = l * Math.ceil(e[1] / 2),
            f = ua()
          return (
            '\n    vec4 ' +
            r +
            '(int b, int row, int col) {\n      vec2 uv = packedUVfrom3D(\n        ' +
            u +
            ', ' +
            c +
            ', ' +
            h +
            ', ' +
            l +
            ', b, row, col);\n      return ' +
            f.texture2D +
            '(' +
            n +
            ', uv);\n    }\n  '
          )
        })(t)
      default:
        return (function (t) {
          for (
            var e = t.shapeInfo.logicalShape,
              n = e.length,
              r = t.name,
              o = 'get' + r.charAt(0).toUpperCase() + r.slice(1),
              a = t.shapeInfo.texShape,
              i = [Math.ceil(a[0] / 2), Math.ceil(a[1] / 2)],
              s = i[0],
              u = i[1],
              c = Math.ceil(e[n - 1] / 2),
              l = c * Math.ceil(e[n - 2] / 2),
              h = 'int b, int row, int col',
              f = 'b * ' + l + ' + (row / 2) * ' + c + ' + (col / 2)',
              d = 2;
            d < n - 1;
            d++
          )
            (h = 'int b' + d + ', ' + h),
              (l *= e[n - d - 1]),
              (f = 'b' + d + ' * ' + l + ' + ' + f)
          var p = ua()
          return (
            '\n    vec4 ' +
            o +
            '(' +
            h +
            ') {\n      int index = ' +
            f +
            ';\n      int texR = index / ' +
            u +
            ';\n      int texC = index - texR * ' +
            u +
            ';\n      vec2 uv = (vec2(texC, texR) + halfCR) / vec2(' +
            u +
            ', ' +
            s +
            ');\n      return ' +
            p.texture2D +
            '(' +
            r +
            ', uv);\n    }\n  '
          )
        })(t)
    }
  }
  var va =
      '\nvec2 uvFromFlat(int texNumR, int texNumC, int index) {\n  int texR = index / texNumC;\n  int texC = index - texR * texNumC;\n  return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);\n}\nvec2 packedUVfrom1D(int texNumR, int texNumC, int index) {\n  int texelIndex = index / 2;\n  int texR = texelIndex / texNumC;\n  int texC = texelIndex - texR * texNumC;\n  return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);\n}\n',
    ma =
      '\nvec2 packedUVfrom2D(int texelsInLogicalRow, int texNumR,\n  int texNumC, int row, int col) {\n  int texelIndex = (row / 2) * texelsInLogicalRow + (col / 2);\n  int texR = texelIndex / texNumC;\n  int texC = texelIndex - texR * texNumC;\n  return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);\n}\n',
    ga =
      '\nvec2 packedUVfrom3D(int texNumR, int texNumC,\n    int texelsInBatch, int texelsInLogicalRow, int b,\n    int row, int col) {\n  int index = b * texelsInBatch + (row / 2) * texelsInLogicalRow + (col / 2);\n  int texR = index / texNumC;\n  int texC = index - texR * texNumC;\n  return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);\n}\n',
    ya =
      '\n  float getChannel(vec4 frag, vec2 innerDims) {\n    vec2 modCoord = mod(innerDims, 2.);\n    return modCoord.x == 0. ?\n      (modCoord.y == 0. ? frag.r : frag.g) :\n      (modCoord.y == 0. ? frag.b : frag.a);\n  }\n  float getChannel(vec4 frag, int dim) {\n    float modCoord = mod(float(dim), 2.);\n    return modCoord == 0. ? frag.r : frag.g;\n  }\n'
  function xa(t) {
    return 'offset' + t
  }
  function ba(t) {
    var e = t.name,
      n = k(t.shapeInfo.logicalShape)
    return n < 2
      ? 'return ' + e + ';'
      : '\n    for (int i = 0; i < ' +
          n +
          '; i++) {\n      if (i == index) {\n        return ' +
          e +
          '[i];\n      }\n    }\n  '
  }
  function wa(t) {
    if (t <= 1) return 'int'
    if (2 === t) return 'ivec2'
    if (3 === t) return 'ivec3'
    if (4 === t) return 'ivec4'
    if (5 === t) return 'ivec5'
    if (6 === t) return 'ivec6'
    throw Error('GPU for rank ' + t + ' is not yet supported')
  }
  function Ca(t, e) {
    var n = JSON.parse(JSON.stringify(t))
    return (n.shapeInfo.logicalShape = e), n
  }
  function Ea(t, e) {
    return e
      .map(function (e) {
        return t[e]
      })
      .join(', ')
  }
  var Ra = function (t, e, n, r) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        C(t.length > 2, function () {
          return (
            'Packed arg' +
            (n.charAt(0).toUpperCase() + n.slice(1)) +
            ' supports only inputs with rank above 2.'
          )
        })
      var o = t[t.length - 1],
        a = Math.ceil(o / e)
      ;(this.outputShape = t.slice(0, -1)),
        a > 1 && this.outputShape.push(a),
        r || this.variableNames.push('bestIndicesA')
      var i,
        s,
        u = this.outputShape,
        c = u.length,
        l = wa(c),
        h = sa('coords', c)
      if (1 === a) {
        var f = wa((s = c + 1))
        i =
          '\n        ' +
          f +
          ' sourceLocR = ' +
          f +
          '(' +
          h.join() +
          ', 0);\n        ++' +
          h[c - 1] +
          ';\n        ' +
          f +
          ' sourceLocG = ' +
          f +
          '(' +
          h.join() +
          ', 0);\n        ++' +
          h[c - 2] +
          ';\n        ' +
          f +
          ' sourceLocA = ' +
          f +
          '(' +
          h.join() +
          ', 0);\n        --' +
          h[c - 1] +
          ';\n        ' +
          f +
          ' sourceLocB = ' +
          f +
          '(' +
          h.join() +
          ', 0);\n        --' +
          h[c - 2] +
          ';'
      } else
        (s = c),
          (i =
            '\n        ' +
            l +
            ' sourceLocR = coords;\n        ++' +
            h[c - 1] +
            ';\n        ' +
            l +
            ' sourceLocG = coords;\n        ++' +
            h[c - 2] +
            ';\n        ' +
            l +
            ' sourceLocA = coords;\n        --' +
            h[c - 1] +
            ';\n        ' +
            l +
            ' sourceLocB = coords;\n        --' +
            h[c - 2] +
            ';')
      var d = ['x', 'y', 'z', 'w', 'u', 'v'].slice(0, s),
        p = '.' + d[s - 1],
        v = d.map(function (t) {
          return 'int ' + t
        }),
        m = sa('sourceLocR', s - 1).concat('inIdx.r'),
        g = sa('sourceLocG', s - 1).concat('inIdx.g'),
        y = sa('sourceLocB', s - 1).concat('inIdx.b'),
        x = sa('sourceLocA', s - 1).concat('inIdx.a'),
        b = 'max' === n ? 'greaterThan' : 'lessThan',
        w = r
          ? ''
          : '\n          inIdx = round(vec4(getBestIndicesAChannel(' +
            m.join() +
            '),\n                             getBestIndicesAChannel(' +
            g.join() +
            '),\n                             getBestIndicesAChannel(' +
            y.join() +
            '),\n                             getBestIndicesAChannel(' +
            x.join() +
            ')));',
        E =
          'vec4(\n            getAChannel(' +
          m.join() +
          '),\n            hasNextCol ? getAChannel(' +
          g.join() +
          ') : 0.,\n            hasNextRow ? getAChannel(' +
          y.join() +
          ') : 0.,\n            hasNextRow && hasNextCol ? getAChannel(' +
          x.join() +
          ') : 0.)',
        R = r
          ? ''
          : '\n      float getBestIndicesAChannel(' +
            v.join() +
            ') {\n        return getChannel(getBestIndicesA(' +
            d.join() +
            '),\n                                          vec2(' +
            d.slice(-2).join() +
            '));\n      }'
      this.userCode =
        '\n      float getAChannel(' +
        v.join() +
        ') {\n        return getChannel(getA(' +
        d.join() +
        '),\n                               vec2(' +
        d.slice(-2).join() +
        '));\n      }\n      ' +
        R +
        '\n      void main() {\n        ' +
        l +
        ' coords = getOutputCoords();\n        bool hasNextCol = ' +
        h[c - 1] +
        ' < ' +
        (u[c - 1] - 1) +
        ';\n        bool hasNextRow = ' +
        h[c - 2] +
        ' < ' +
        (u[c - 2] - 1) +
        ';\n        ' +
        i +
        '\n        ivec4 srcIdx = ivec4(sourceLocR' +
        p +
        ', sourceLocG' +
        p +
        ',\n          sourceLocB' +
        p +
        ', sourceLocA' +
        p +
        ') * ' +
        e +
        ';\n        ivec4 inIdx = srcIdx;\n        vec4 bestIndex = vec4(inIdx);\n        vec4 bestValue = ' +
        E +
        ';\n\n        for (int i = 0; i < ' +
        e +
        '; i++) {\n          inIdx = srcIdx;\n          ' +
        w +
        '\n          vec4 candidate = ' +
        E +
        ';\n          bvec4 nan = isnan(candidate);\n          bvec4 replace = bvec4(\n            vec4(' +
        b +
        '(candidate, bestValue)) * (vec4(1.0) - vec4(nan)));\n\n          bestValue = vec4(replace.x  ? candidate.x : bestValue.x,\n                           replace.y  ? candidate.y : bestValue.y,\n                           replace.z  ? candidate.z : bestValue.z,\n                           replace.w  ? candidate.w : bestValue.w);\n          bestIndex = mix(bestIndex, vec4(inIdx), vec4(replace));\n          srcIdx++;\n        }\n        setOutput(bestIndex);\n      }\n    '
    },
    Ia = function (t) {
      ;(this.variableNames = ['dy']), (this.outputShape = t.inShape)
      var e = t.filterHeight,
        n = t.filterWidth,
        r = t.strideHeight,
        o = t.strideWidth,
        a = t.dilationHeight,
        i = t.dilationWidth,
        s = t.effectiveFilterHeight,
        u = t.effectiveFilterWidth,
        c = s - 1 - t.padInfo.top,
        l = u - 1 - t.padInfo.left,
        h = 1 / (e * n)
      this.userCode =
        '\n      const ivec2 pads = ivec2(' +
        c +
        ', ' +
        l +
        ');\n      const float avgMultiplier = float(' +
        h +
        ');\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n\n        ivec2 dyRCCorner = coords.yz - pads;\n        int dyRCorner = dyRCCorner.x;\n        int dyCCorner = dyRCCorner.y;\n\n        // Convolve dy(?, ?, d) with pos mask(:, :, d) to get dx(xR, xC, d).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n        for (int wR = 0; wR < ' +
        s +
        ';\n            wR += ' +
        a +
        ') {\n          float dyR = float(dyRCorner + wR) / ' +
        r +
        '.0;\n\n          if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 || fract(dyR) > 0.0) {\n            continue;\n          }\n          int idyR = int(dyR);\n\n          for (int wC = 0; wC < ' +
        u +
        ';\n            wC+= ' +
        i +
        ') {\n            float dyC = float(dyCCorner + wC) / ' +
        o +
        '.0;\n\n            if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                fract(dyC) > 0.0) {\n              continue;\n            }\n            int idyC = int(dyC);\n\n            float dyValue = getDy(b, idyR, idyC, d);\n\n            dotProd += dyValue * avgMultiplier;\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    ka = function (t) {
      ;(this.variableNames = ['dy']), (this.outputShape = t.inShape)
      var e = t.filterDepth,
        n = t.filterHeight,
        r = t.filterWidth,
        o = t.strideDepth,
        a = t.strideHeight,
        i = t.strideWidth,
        s = t.dilationDepth,
        u = t.dilationHeight,
        c = t.dilationWidth,
        l = t.effectiveFilterDepth,
        h = t.effectiveFilterHeight,
        f = t.effectiveFilterWidth,
        d = l - 1 - t.padInfo.front,
        p = h - 1 - t.padInfo.top,
        v = f - 1 - t.padInfo.left,
        m = 1 / (e * n * r)
      this.userCode =
        '\n      const ivec3 pads = ivec3(' +
        d +
        ', ' +
        p +
        ', ' +
        v +
        ');\n      const float avgMultiplier = float(' +
        m +
        ');\n\n      void main() {\n        ivec5 coords = getOutputCoords();\n        int batch = coords.x;\n        int ch = coords.u;\n\n        ivec3 dyCorner = ivec3(coords.y, coords.z, coords.w) - pads;\n        int dyDCorner = dyCorner.x;\n        int dyRCorner = dyCorner.y;\n        int dyCCorner = dyCorner.z;\n\n        // Convolve dy(?, ?, ?, d) with pos mask(:, :, :, ch) to get\n        // dx(xD, xR, xC, ch).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n\n        for (int wD = 0; wD < ' +
        l +
        ';\n            wD += ' +
        s +
        ') {\n          float dyD = float(dyDCorner + wD) / ' +
        o +
        '.0;\n\n          if (dyD < 0.0 || dyD >= ' +
        t.outDepth +
        '.0 || fract(dyD) > 0.0) {\n            continue;\n          }\n          int idyD = int(dyD);\n\n          for (int wR = 0; wR < ' +
        h +
        ';\n              wR += ' +
        u +
        ') {\n            float dyR = float(dyRCorner + wR) / ' +
        a +
        '.0;\n\n            if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 ||\n                fract(dyR) > 0.0) {\n              continue;\n            }\n            int idyR = int(dyR);\n\n            for (int wC = 0; wC < ' +
        f +
        ';\n                wC += ' +
        c +
        ') {\n              float dyC = float(dyCCorner + wC) / ' +
        i +
        '.0;\n\n              if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                  fract(dyC) > 0.0) {\n                continue;\n              }\n              int idyC = int(dyC);\n\n              float dyValue = getDy(batch, idyD, idyR, idyC, ch);\n\n              dotProd += dyValue * avgMultiplier;\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Sa = function (t, e, n, r, o, a) {
      ;(this.outputShape = []),
        (this.variableNames = ['x', 'mean', 'variance']),
        Ro(t, e),
        Ro(t, n)
      var i = '0.0'
      null != r &&
        (Ro(t, r),
        this.variableNames.push('offset'),
        (i = 'getOffsetAtOutCoords()'))
      var s = '1.0'
      null != o &&
        (Ro(t, o),
        this.variableNames.push('scale'),
        (s = 'getScaleAtOutCoords()')),
        (this.outputShape = t),
        (this.userCode =
          '\n      void main() {\n        float x = getXAtOutCoords();\n        float mean = getMeanAtOutCoords();\n        float variance = getVarianceAtOutCoords();\n        float offset = ' +
          i +
          ';\n        float scale = ' +
          s +
          ';\n        float inv = scale * inversesqrt(variance + float(' +
          a +
          '));\n        setOutput(dot(vec3(x, -mean, offset), vec3(inv, inv, 1)));\n      }\n    ')
    },
    Aa = function (t, e, n, r, o, a) {
      ;(this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.variableNames = ['x', 'mean', 'variance']),
        Ro(t, e),
        Ro(t, n)
      var i = 'vec4(0.0)'
      null != r &&
        (Ro(t, r),
        this.variableNames.push('offset'),
        (i = 'getOffsetAtOutCoords()'))
      var s = 'vec4(1.0)'
      null != o &&
        (Ro(t, o),
        this.variableNames.push('scale'),
        (s = 'getScaleAtOutCoords()')),
        (this.outputShape = t),
        (this.userCode =
          '\n      void main() {\n        vec4 offset = ' +
          i +
          ';\n        vec4 scale = ' +
          s +
          ';\n\n        vec4 x = getXAtOutCoords();\n        vec4 mean = getMeanAtOutCoords();\n        vec4 variance = getVarianceAtOutCoords();\n\n        vec4 inv = scale * inversesqrt(variance + vec4(' +
          a +
          '));\n\n        setOutput((x - mean) * inv + offset);\n      }\n    ')
    },
    Da = 'return areal * breal - aimag * bimag;',
    Ta = 'return areal * bimag + aimag * breal;',
    Na = function (t, e, n) {
      ;(this.variableNames = ['AReal', 'AImag', 'BReal', 'BImag']),
        (this.outputShape = Ro(e, n)),
        (this.userCode =
          '\n      float binaryOpComplex(\n          float areal, float aimag, float breal, float bimag) {\n        ' +
          t +
          '\n      }\n\n      void main() {\n        float areal = getARealAtOutCoords();\n        float aimag = getAImagAtOutCoords();\n        float breal = getBRealAtOutCoords();\n        float bimag = getBImagAtOutCoords();\n        setOutput(binaryOpComplex(areal, aimag, breal, bimag));\n      }\n    ')
    },
    Fa = 'return a + b;',
    _a = 'return a - b;',
    Oa = 'return a * b;',
    Ma = 'return (a < 0.) ? b * a : a;',
    Ba = function (t, e, n) {
      ;(this.variableNames = ['A', 'B']),
        (this.outputShape = Ro(e, n)),
        (this.userCode =
          '\n      float binaryOperation(float a, float b) {\n        ' +
          t +
          '\n      }\n\n      void main() {\n        float a = getAAtOutCoords();\n        float b = getBAtOutCoords();\n        setOutput(binaryOperation(a, b));\n      }\n    ')
    },
    Pa =
      '\n  vec4 aLessThanZero = vec4(lessThan(a, vec4(0.)));\n  return (aLessThanZero * (b * a)) + ((vec4(1.0) - aLessThanZero) * a);\n',
    La = function (t, e, n, r) {
      void 0 === r && (r = !1),
        (this.variableNames = ['A', 'B']),
        (this.supportsBroadcasting = !0),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = Ro(e, n))
      var o = this.outputShape.length,
        a = ''
      if (r)
        if (0 === o || 1 === k(this.outputShape))
          a =
            '\n          result.y = 0.;\n          result.z = 0.;\n          result.w = 0.;\n        '
        else if (
          ((a =
            '\n          ' + wa(o) + ' coords = getOutputCoords();\n        '),
          1 === o)
        )
          a +=
            '\n            result.y = (coords + 1) >= ' +
            this.outputShape[0] +
            ' ? 0. : result.y;\n            result.z = 0.;\n            result.w = 0.;\n          '
        else {
          var i = sa('coords', o)
          a +=
            '\n            bool nextRowOutOfBounds =\n              (' +
            i[o - 2] +
            ' + 1) >= ' +
            this.outputShape[o - 2] +
            ';\n            bool nextColOutOfBounds =\n              (' +
            i[o - 1] +
            ' + 1) >= ' +
            this.outputShape[o - 1] +
            ';\n            result.y = nextColOutOfBounds ? 0. : result.y;\n            result.z = nextRowOutOfBounds ? 0. : result.z;\n            result.w = nextColOutOfBounds || nextRowOutOfBounds ? 0. : result.w;\n          '
        }
      this.userCode =
        '\n      vec4 binaryOperation(vec4 a, vec4 b) {\n        ' +
        t +
        '\n      }\n\n      void main() {\n        vec4 a = getAAtOutCoords();\n        vec4 b = getBAtOutCoords();\n\n        vec4 result = binaryOperation(a, b);\n        ' +
        a +
        '\n\n        setOutput(result);\n      }\n    '
    },
    Wa = (function () {
      function t(t) {
        ;(this.variableNames = ['A']),
          (this.outputShape = t),
          (this.userCode =
            '\n      uniform float minVal;\n      uniform float maxVal;\n\n      void main() {\n        float value = getAAtOutCoords();\n        if (isnan(value)) {\n          setOutput(value);\n          return;\n        }\n\n        setOutput(clamp(value, minVal, maxVal));\n      }\n    ')
      }
      return (
        (t.prototype.getCustomSetupFunc = function (t, e) {
          var n = this
          return function (r, o) {
            null == n.minLoc &&
              ((n.minLoc = r.getUniformLocationNoThrow(o, 'minVal')),
              (n.maxLoc = r.getUniformLocationNoThrow(o, 'maxVal'))),
              r.gl.uniform1f(n.minLoc, t),
              r.gl.uniform1f(n.maxLoc, e)
          }
        }),
        t
      )
    })(),
    Ua = (function () {
      function t(t) {
        ;(this.variableNames = ['A']),
          (this.packedInputs = !0),
          (this.packedOutput = !0),
          (this.outputShape = t),
          (this.userCode =
            '\n      uniform float minVal;\n      uniform float maxVal;\n\n      void main() {\n        vec4 value = getAAtOutCoords();\n\n        if (any(isnan(value))) {\n          setOutput(value);\n          return;\n        }\n\n        setOutput(clamp(value, vec4(minVal), vec4(maxVal)));\n      }\n    ')
      }
      return (
        (t.prototype.getCustomSetupFunc = function (t, e) {
          var n = this
          return function (r, o) {
            null == n.minLoc &&
              ((n.minLoc = r.getUniformLocationNoThrow(o, 'minVal')),
              (n.maxLoc = r.getUniformLocationNoThrow(o, 'maxVal'))),
              r.gl.uniform1f(n.minLoc, t),
              r.gl.uniform1f(n.maxLoc, e)
          }
        }),
        t
      )
    })(),
    Va = function (t) {
      ;(this.variableNames = ['real', 'imag']),
        (this.outputShape = t),
        (this.userCode =
          '\n      void main() {\n        float re = abs(getRealAtOutCoords());\n        float im = abs(getImagAtOutCoords());\n        float mx = max(re, im);\n\n        // sadly the length function in glsl is not underflow-safe\n        // (at least not on Intel GPUs). So the safe solution is\n        // to ensure underflow-safety in all cases.\n        setOutput(\n          mx == 0.0 ? 0.0 : mx * length(vec2(1, min(re, im)/mx))\n        );\n      }\n    ')
    },
    za = function (t) {
      ;(this.outputShape = []),
        (this.outputShape = Sn(t, 1)),
        (this.variableNames = t.map(function (t, e) {
          return 'T' + e
        }))
      var e = new Array(t.length - 1)
      e[0] = t[0][1]
      for (var n = 1; n < e.length; n++) e[n] = e[n - 1] + t[n][1]
      var r = ['if (yC < ' + e[0] + ') setOutput(getT0(yR, yC));']
      for (n = 1; n < e.length; n++) {
        var o = e[n - 1]
        r.push(
          'else if (yC < ' +
            e[n] +
            ') setOutput(getT' +
            n +
            '(yR, yC-' +
            o +
            '));'
        )
      }
      var a = e.length,
        i = e[e.length - 1]
      r.push('else setOutput(getT' + a + '(yR, yC-' + i + '));'),
        (this.userCode =
          '\n      void main() {\n        ivec2 coords = getOutputCoords();\n        int yR = coords.x;\n        int yC = coords.y;\n\n        ' +
          r.join('\n        ') +
          '\n      }\n    ')
    },
    Ga = function (t, e) {
      ;(this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = []),
        (this.outputShape = Sn(t, e))
      var n = this.outputShape,
        r = n.length,
        o = wa(r),
        a = sa('coords', r),
        i = ['x', 'y', 'z', 'w', 'u', 'v'].slice(0, r)
      this.variableNames = t.map(function (t, e) {
        return 'T' + e
      })
      var s = new Array(t.length - 1)
      s[0] = t[0][e]
      for (var u = 1; u < s.length; u++) s[u] = s[u - 1] + t[u][e]
      var c = i[e],
        l = i.slice(-2),
        h = i.join(),
        f =
          'if (' +
          c +
          ' < ' +
          s[0] +
          ') {\n        return getChannel(\n            getT0(' +
          h +
          '), vec2(' +
          l.join() +
          '));\n        }'
      for (u = 1; u < s.length; u++) {
        var d = s[u - 1]
        f +=
          '\n        if (' +
          c +
          ' < ' +
          s[u] +
          '  && ' +
          c +
          ' >= ' +
          s[u - 1] +
          ') {\n          return getChannel(\n            getT' +
          u +
          '(' +
          Ha(i, c, d) +
          '),\n            vec2(' +
          Ha(l, c, d) +
          '));\n        }'
      }
      var p = s.length,
        v = s[s.length - 1]
      ;(f +=
        '\n        return getChannel(\n          getT' +
        p +
        '(' +
        Ha(i, c, v) +
        '),\n          vec2(' +
        Ha(l, c, v) +
        '));'),
        (this.userCode =
          '\n      float getValue(' +
          i.map(function (t) {
            return 'int ' + t
          }) +
          ') {\n        ' +
          f +
          '\n      }\n\n      void main() {\n        ' +
          o +
          ' coords = getOutputCoords();\n        vec4 result = vec4(getValue(' +
          a +
          '), 0., 0., 0.);\n\n        ' +
          a[r - 1] +
          ' = ' +
          a[r - 1] +
          ' + 1;\n        if (' +
          a[r - 1] +
          ' < ' +
          n[r - 1] +
          ') {\n          result.g = getValue(' +
          a +
          ');\n        }\n\n        ' +
          a[r - 2] +
          ' = ' +
          a[r - 2] +
          ' + 1;\n        if (' +
          a[r - 2] +
          ' < ' +
          n[r - 2] +
          ') {\n          result.a = getValue(' +
          a +
          ');\n        }\n\n        ' +
          a[r - 1] +
          ' = ' +
          a[r - 1] +
          ' - 1;\n        if (' +
          a[r - 2] +
          ' < ' +
          n[r - 2] +
          ' &&\n            ' +
          a[r - 1] +
          ' < ' +
          n[r - 1] +
          ') {\n          result.b = getValue(' +
          a +
          ');\n        }\n        setOutput(result);\n      }\n    ')
    }
  function Ha(t, e, n) {
    var r = t.indexOf(e)
    return t
      .map(function (t, e) {
        return e === r ? t + ' - ' + n : t
      })
      .join()
  }
  var qa = function (t) {
      ;(this.variableNames = ['x', 'dy']), (this.outputShape = t.filterShape)
      var e = t.strideHeight,
        n = t.strideWidth,
        r = t.padInfo.top,
        o = t.padInfo.left,
        a = 'channelsLast' === t.dataFormat
      this.userCode =
        '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int wR = coords.x;\n        int wC = coords.y;\n        int d1 = coords.z;\n        int d2 = coords.w;\n\n        // Convolve x(?, ?, d1) with dy(:, :, d2) to get dw(wR, wC, d1, d2).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n\n        for (int b = 0; b < ' +
        t.batchSize +
        '; b++) {\n          for (int yR = 0; yR < ' +
        t.outHeight +
        '; yR++) {\n            int xR = wR + yR * ' +
        e +
        ' - ' +
        r +
        ';\n\n            if (xR < 0 || xR >= ' +
        t.inHeight +
        ') {\n              continue;\n            }\n\n            for (int yC = 0; yC < ' +
        t.outWidth +
        '; yC++) {\n              int xC = wC + yC * ' +
        n +
        ' - ' +
        o +
        ';\n\n              if (xC < 0 || xC >= ' +
        t.inWidth +
        ') {\n                continue;\n              }\n\n              if (' +
        a +
        ') {\n                float dyValue = getDy(b, yR, yC, d2);\n                float xValue = getX(b, xR, xC, d1);\n                dotProd += (xValue * dyValue);\n              } else {\n                float dyValue = getDy(b, d2, yR, yC);\n                float xValue = getX(b, d1, xR, xC);\n                dotProd += (xValue * dyValue);\n              }\n\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Ka = function (t) {
      ;(this.variableNames = ['dy', 'W']), (this.outputShape = t.inShape)
      var e = t.filterHeight,
        n = t.filterWidth,
        r = t.strideHeight,
        o = t.strideWidth,
        a = 'channelsLast' === t.dataFormat,
        i = e - 1 - t.padInfo.top,
        s = n - 1 - t.padInfo.left,
        u = a ? 1 : 2,
        c = a ? 2 : 3,
        l = a ? 3 : 1
      this.userCode =
        '\n      const ivec2 pads = ivec2(' +
        i +
        ', ' +
        s +
        ');\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int batch = coords[0];\n        int d1 = coords[' +
        l +
        '];\n\n        ivec2 dyCorner = ivec2(coords[' +
        u +
        '], coords[' +
        c +
        ']) - pads;\n        int dyRCorner = dyCorner.x;\n        int dyCCorner = dyCorner.y;\n\n        // Convolve dy(?, ?, d2) with w(:, :, d1, d2) to compute dx(xR, xC, d1).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n        for (int wR = 0; wR < ' +
        e +
        '; wR++) {\n          float dyR = float(dyRCorner + wR) / ' +
        r +
        '.0;\n\n          if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 || fract(dyR) > 0.0) {\n            continue;\n          }\n          int idyR = int(dyR);\n\n          int wRPerm = ' +
        e +
        ' - 1 - wR;\n\n          for (int wC = 0; wC < ' +
        n +
        '; wC++) {\n            float dyC = float(dyCCorner + wC) / ' +
        o +
        '.0;\n\n            if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                fract(dyC) > 0.0) {\n              continue;\n            }\n            int idyC = int(dyC);\n\n            int wCPerm = ' +
        n +
        ' - 1 - wC;\n\n            for (int d2 = 0; d2 < ' +
        t.outChannels +
        '; d2++) {\n\n              if (' +
        a +
        ') {\n                float xValue = getDy(batch, idyR, idyC, d2);\n                float wValue = getW(wRPerm, wCPerm, d1, d2);\n                dotProd += xValue * wValue;\n              } else {\n                float xValue = getDy(batch, d2, idyR, idyC);\n                float wValue = getW(wRPerm, wCPerm, d1, d2);\n                dotProd += xValue * wValue;\n              }\n\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    ja = function (t) {
      ;(this.variableNames = ['x', 'dy']), (this.outputShape = t.filterShape)
      var e = t.strideDepth,
        n = t.strideHeight,
        r = t.strideWidth,
        o = t.padInfo.front,
        a = t.padInfo.top,
        i = t.padInfo.left
      this.userCode =
        '\n      void main() {\n        ivec5 coords = getOutputCoords();\n        int wF = coords.x;\n        int wR = coords.y;\n        int wC = coords.z;\n        int d1 = coords.w;\n        int d2 = coords.u;\n\n        float dotProd = 0.0;\n\n        for (int b = 0; b < ' +
        t.batchSize +
        '; b++) {\n          for (int yF = 0; yF < ' +
        t.outDepth +
        '; yF++) {\n            int xF = wF + yF * ' +
        e +
        ' - ' +
        o +
        ';\n\n            if (xF < 0 || xF >= ' +
        t.inDepth +
        ') {\n              continue;\n            }\n\n            for (int yR = 0; yR < ' +
        t.outHeight +
        '; yR++) {\n              int xR = wR + yR * ' +
        n +
        ' - ' +
        a +
        ';\n\n              if (xR < 0 || xR >= ' +
        t.inHeight +
        ') {\n                continue;\n              }\n\n              for (int yC = 0; yC < ' +
        t.outWidth +
        '; yC++) {\n                int xC = wC + yC * ' +
        r +
        ' - ' +
        i +
        ';\n\n                if (xC < 0 || xC >= ' +
        t.inWidth +
        ') {\n                  continue;\n                }\n\n                float dyValue = getDy(b, yF, yR, yC, d2);\n                float xValue = getX(b, xF, xR, xC, d1);\n                dotProd += (xValue * dyValue);\n              }\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Xa = function (t) {
      ;(this.variableNames = ['dy', 'W']), (this.outputShape = t.inShape)
      var e = t.filterDepth,
        n = t.filterHeight,
        r = t.filterWidth,
        o = t.strideDepth,
        a = t.strideHeight,
        i = t.strideWidth,
        s = e - 1 - t.padInfo.front,
        u = n - 1 - t.padInfo.top,
        c = r - 1 - t.padInfo.left
      this.userCode =
        '\n      const ivec3 pads = ivec3(' +
        s +
        ', ' +
        u +
        ', ' +
        c +
        ');\n\n      void main() {\n        ivec5 coords = getOutputCoords();\n        int batch = coords.x;\n        int d1 = coords.u;\n\n\n        ivec3 dyCorner = ivec3(coords.y, coords.z, coords.w) - pads;\n        int dyFCorner = dyCorner.x;\n        int dyRCorner = dyCorner.y;\n        int dyCCorner = dyCorner.z;\n\n        float dotProd = 0.0;\n        for (int wF = 0; wF < ' +
        e +
        '; wF++) {\n          float dyF = float(dyFCorner + wF) / ' +
        o +
        '.0;\n\n          if (dyF < 0.0 || dyF >= ' +
        t.outDepth +
        '.0 || fract(dyF) > 0.0) {\n            continue;\n          }\n          int idyF = int(dyF);\n\n          int wFPerm = ' +
        e +
        ' - 1 - wF;\n\n          for (int wR = 0; wR < ' +
        n +
        '; wR++) {\n            float dyR = float(dyRCorner + wR) / ' +
        a +
        '.0;\n\n            if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 ||\n              fract(dyR) > 0.0) {\n              continue;\n            }\n            int idyR = int(dyR);\n\n            int wRPerm = ' +
        n +
        ' - 1 - wR;\n\n            for (int wC = 0; wC < ' +
        r +
        '; wC++) {\n              float dyC = float(dyCCorner + wC) / ' +
        i +
        '.0;\n\n              if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                  fract(dyC) > 0.0) {\n                continue;\n              }\n              int idyC = int(dyC);\n\n              int wCPerm = ' +
        r +
        ' - 1 - wC;\n\n              for (int d2 = 0; d2 < ' +
        t.outChannels +
        '; d2++) {\n                float xValue = getDy(batch, idyF, idyR, idyC, d2);\n                float wValue = getW(wFPerm, wRPerm, wCPerm, d1, d2);\n                dotProd += xValue * wValue;\n              }\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Ya = function (t) {
      ;(this.variableNames = ['x', 'dy']), (this.outputShape = t.filterShape)
      var e = t.strideHeight,
        n = t.strideWidth,
        r = t.padInfo.top,
        o = t.padInfo.left,
        a = t.outChannels / t.inChannels
      this.userCode =
        '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int wR = coords.x;\n        int wC = coords.y;\n        int d1 = coords.z;\n        int dm = coords.w;\n        int d2 = d1 * ' +
        a +
        ' + dm;\n\n        float dotProd = 0.0;\n\n        // TO DO: Vec4 over the batch size\n        for (int b = 0; b < ' +
        t.batchSize +
        '; b++) {\n          for (int yR = 0; yR < ' +
        t.outHeight +
        '; yR++) {\n            int xR = wR + yR * ' +
        e +
        ' - ' +
        r +
        ';\n\n            if (xR < 0 || xR >= ' +
        t.inHeight +
        ') {\n              continue;\n            }\n\n            for (int yC = 0; yC < ' +
        t.outWidth +
        '; yC++) {\n              int xC = wC + yC * ' +
        n +
        ' - ' +
        o +
        ';\n\n              if (xC < 0 || xC >= ' +
        t.inWidth +
        ') {\n                continue;\n              }\n\n              float dyValue = getDy(b, yR, yC, d2);\n              float xValue = getX(b, xR, xC, d1);\n              dotProd += (xValue * dyValue);\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    $a = function (t) {
      ;(this.variableNames = ['dy', 'W']), (this.outputShape = t.inShape)
      var e = t.filterHeight,
        n = t.filterWidth,
        r = t.strideHeight,
        o = t.strideWidth,
        a = e - 1 - t.padInfo.top,
        i = n - 1 - t.padInfo.left,
        s = t.outChannels / t.inChannels
      this.userCode =
        '\n      const ivec2 pads = ivec2(' +
        a +
        ', ' +
        i +
        ');\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int batch = coords[0];\n        int d1 = coords[3];\n        ivec2 dyCorner = coords.yz - pads;\n        int dyRCorner = dyCorner.x;\n        int dyCCorner = dyCorner.y;\n\n        float dotProd = 0.0;\n\n        for (int wR = 0; wR < ' +
        e +
        '; wR++) {\n          float dyR = float(dyRCorner + wR) / ' +
        r +
        '.0;\n\n          if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 || fract(dyR) > 0.0) {\n            continue;\n          }\n          int idyR = int(dyR);\n\n          int wRPerm = ' +
        e +
        ' - 1 - wR;\n\n          for (int wC = 0; wC < ' +
        n +
        '; wC++) {\n            float dyC = float(dyCCorner + wC) / ' +
        o +
        '.0;\n\n            if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                fract(dyC) > 0.0) {\n              continue;\n            }\n            int idyC = int(dyC);\n\n            int wCPerm = ' +
        n +
        ' - 1 - wC;\n\n            // TO DO: Vec4 over the channelMul\n            for (int dm = 0; dm < ' +
        s +
        '; dm++) {\n              int d2 = d1 * ' +
        s +
        ' + dm;\n              float xValue = getDy(batch, idyR, idyC, d2);\n              float wValue = getW(wRPerm, wCPerm, d1, dm);\n              dotProd += xValue * wValue;\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Qa = function (t, e, n, r) {
      void 0 === e && (e = !1),
        void 0 === n && (n = null),
        void 0 === r && (r = !1),
        (this.variableNames = ['x', 'W']),
        (this.outputShape = t.outShape)
      var o = t.padInfo.top,
        a = t.padInfo.left,
        i = t.strideHeight,
        s = t.strideWidth,
        u = t.dilationHeight,
        c = t.dilationWidth,
        l = t.filterHeight,
        h = t.filterWidth,
        f = 4 * Math.floor(t.inChannels / 4),
        d = t.inChannels % 4,
        p = 'channelsLast' === t.dataFormat,
        v = p ? 1 : 2,
        m = p ? 2 : 3,
        g = p ? 3 : 1,
        y = '',
        x = ''
      n &&
        ((y = r
          ? 'float activation(float a) {\n          float b = getPreluActivationWeightsAtOutCoords();\n          ' +
            n +
            '\n        }'
          : '\n          float activation(float x) {\n            ' +
            n +
            '\n          }\n        '),
        (x = 'result = activation(result);'))
      var b = e ? 'result += getBiasAtOutCoords();' : ''
      e && this.variableNames.push('bias'),
        r && this.variableNames.push('preluActivationWeights'),
        (this.userCode =
          '\n      ' +
          y +
          '\n\n      const ivec2 strides = ivec2(' +
          i +
          ', ' +
          s +
          ');\n      const ivec2 pads = ivec2(' +
          o +
          ', ' +
          a +
          ');\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int batch = coords[0];\n        int d2 = coords[' +
          g +
          '];\n\n        ivec2 xRCCorner =\n            ivec2(coords[' +
          v +
          '], coords[' +
          m +
          ']) * strides - pads;\n        int xRCorner = xRCCorner.x;\n        int xCCorner = xRCCorner.y;\n\n        // Convolve x(?, ?, d1) with w(:, :, d1, d2) to get y(yR, yC, d2).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n        for (int wR = 0; wR < ' +
          l +
          '; wR++) {\n          int xR = xRCorner + wR * ' +
          u +
          ';\n\n          if (xR < 0 || xR >= ' +
          t.inHeight +
          ') {\n            continue;\n          }\n\n          for (int wC = 0; wC < ' +
          h +
          '; wC++) {\n            int xC = xCCorner + wC * ' +
          c +
          ';\n\n            if (xC < 0 || xC >= ' +
          t.inWidth +
          ') {\n              continue;\n            }\n\n            for (int d1 = 0; d1 < ' +
          f +
          '; d1 += 4) {\n              vec4 wValues = vec4(\n                getW(wR, wC, d1, d2),\n                getW(wR, wC, d1 + 1, d2),\n                getW(wR, wC, d1 + 2, d2),\n                getW(wR, wC, d1 + 3, d2)\n              );\n\n              if (' +
          p +
          ') {\n                vec4 xValues = vec4(\n                  getX(batch, xR, xC, d1),\n                  getX(batch, xR, xC, d1 + 1),\n                  getX(batch, xR, xC, d1 + 2),\n                  getX(batch, xR, xC, d1 + 3)\n                );\n                dotProd += dot(xValues, wValues);\n              } else {\n                vec4 xValues = vec4(\n                  getX(batch, d1, xR, xC),\n                  getX(batch, d1 + 1, xR, xC),\n                  getX(batch, d1 + 2, xR, xC),\n                  getX(batch, d1 + 3, xR, xC)\n                );\n                dotProd += dot(xValues, wValues);\n              }\n            }\n\n            if (' +
          (1 === d) +
          ') {\n\n              if (' +
          p +
          ') {\n                dotProd +=\n                    getX(batch, xR, xC, ' +
          f +
          ') *\n                    getW(wR, wC, ' +
          f +
          ', d2);\n              } else {\n                dotProd +=\n                    getX(batch, ' +
          f +
          ', xR, xC) *\n                    getW(wR, wC, ' +
          f +
          ', d2);\n              }\n\n            } else if (' +
          (2 === d) +
          ') {\n              vec2 wValues = vec2(\n                getW(wR, wC, ' +
          f +
          ', d2),\n                getW(wR, wC, ' +
          f +
          ' + 1, d2)\n              );\n\n              if (' +
          p +
          ') {\n                vec2 xValues = vec2(\n                  getX(batch, xR, xC, ' +
          f +
          '),\n                  getX(batch, xR, xC, ' +
          f +
          ' + 1)\n                );\n                dotProd += dot(xValues, wValues);\n              } else {\n                vec2 xValues = vec2(\n                  getX(batch, ' +
          f +
          ', xR, xC),\n                  getX(batch, ' +
          f +
          ' + 1, xR, xC)\n                );\n                dotProd += dot(xValues, wValues);\n              }\n\n            } else if (' +
          (3 === d) +
          ') {\n              vec3 wValues = vec3(\n                getW(wR, wC, ' +
          f +
          ', d2),\n                getW(wR, wC, ' +
          f +
          ' + 1, d2),\n                getW(wR, wC, ' +
          f +
          ' + 2, d2)\n              );\n\n              if (' +
          p +
          ') {\n                vec3 xValues = vec3(\n                  getX(batch, xR, xC, ' +
          f +
          '),\n                  getX(batch, xR, xC, ' +
          f +
          ' + 1),\n                  getX(batch, xR, xC, ' +
          f +
          ' + 2)\n                );\n                dotProd += dot(xValues, wValues);\n              } else {\n                vec3 xValues = vec3(\n                  getX(batch, ' +
          f +
          ', xR, xC),\n                  getX(batch, ' +
          f +
          ' + 1, xR, xC),\n                  getX(batch, ' +
          f +
          ' + 2, xR, xC)\n                );\n                dotProd += dot(xValues, wValues);\n              }\n\n            }\n          }\n        }\n\n        float result = dotProd;\n        ' +
          b +
          '\n        ' +
          x +
          '\n        setOutput(result);\n      }\n    ')
    },
    Ja = function (t) {
      ;(this.variableNames = ['x', 'W']), (this.outputShape = t.outShape)
      var e = t.padInfo.front,
        n = t.padInfo.top,
        r = t.padInfo.left,
        o = t.strideDepth,
        a = t.strideHeight,
        i = t.strideWidth,
        s = t.dilationDepth,
        u = t.dilationHeight,
        c = t.dilationWidth,
        l = t.filterDepth,
        h = t.filterHeight,
        f = t.filterWidth,
        d = 4 * Math.floor(t.inChannels / 4),
        p = t.inChannels % 4
      this.userCode =
        '\n      const ivec3 strides = ivec3(' +
        o +
        ', ' +
        a +
        ', ' +
        i +
        ');\n      const ivec3 pads = ivec3(' +
        e +
        ', ' +
        n +
        ', ' +
        r +
        ');\n\n      void main() {\n        ivec5 coords = getOutputCoords();\n        int batch = coords.x;\n        int d2 = coords.u;\n\n        ivec3 xFRCCorner = ivec3(coords.y, coords.z, coords.w) * strides - pads;\n        int xFCorner = xFRCCorner.x;\n        int xRCorner = xFRCCorner.y;\n        int xCCorner = xFRCCorner.z;\n\n        // Convolve x(?, ?, ?, d1) with w(:, :, :, d1, d2) to get\n        // y(yF, yR, yC, d2). ? = to be determined. : = across all\n        // values in that axis.\n        float dotProd = 0.0;\n        for (int wF = 0; wF < ' +
        l +
        '; wF++) {\n          int xF = xFCorner + wF * ' +
        s +
        ';\n\n          if (xF < 0 || xF >= ' +
        t.inDepth +
        ') {\n            continue;\n          }\n\n          for (int wR = 0; wR < ' +
        h +
        '; wR++) {\n            int xR = xRCorner + wR * ' +
        u +
        ';\n\n            if (xR < 0 || xR >= ' +
        t.inHeight +
        ') {\n              continue;\n            }\n\n            for (int wC = 0; wC < ' +
        f +
        '; wC++) {\n              int xC = xCCorner + wC * ' +
        c +
        ';\n\n              if (xC < 0 || xC >= ' +
        t.inWidth +
        ') {\n                continue;\n              }\n\n              for (int d1 = 0; d1 < ' +
        d +
        '; d1 += 4) {\n                vec4 xValues = vec4(\n                  getX(batch, xF, xR, xC, d1),\n                  getX(batch, xF, xR, xC, d1 + 1),\n                  getX(batch, xF, xR, xC, d1 + 2),\n                  getX(batch, xF, xR, xC, d1 + 3)\n                );\n                vec4 wValues = vec4(\n                  getW(wF, wR, wC, d1, d2),\n                  getW(wF, wR, wC, d1 + 1, d2),\n                  getW(wF, wR, wC, d1 + 2, d2),\n                  getW(wF, wR, wC, d1 + 3, d2)\n                );\n\n                dotProd += dot(xValues, wValues);\n              }\n\n              if (' +
        (1 === p) +
        ') {\n                dotProd +=\n                  getX(batch, xF, xR, xC, ' +
        d +
        ') *\n                  getW(wF, wR, wC, ' +
        d +
        ', d2);\n              } else if (' +
        (2 === p) +
        ') {\n                vec2 xValues = vec2(\n                  getX(batch, xF, xR, xC, ' +
        d +
        '),\n                  getX(batch, xF, xR, xC, ' +
        d +
        ' + 1)\n                );\n                vec2 wValues = vec2(\n                  getW(wF, wR, wC, ' +
        d +
        ', d2),\n                  getW(wF, wR, wC, ' +
        d +
        ' + 1, d2)\n                );\n                dotProd += dot(xValues, wValues);\n              } else if (' +
        (3 === p) +
        ') {\n                vec3 xValues = vec3(\n                  getX(batch, xF, xR, xC, ' +
        d +
        '),\n                  getX(batch, xF, xR, xC, ' +
        d +
        ' + 1),\n                  getX(batch, xF, xR, xC, ' +
        d +
        ' + 2)\n                );\n                vec3 wValues = vec3(\n                  getW(wF, wR, wC, ' +
        d +
        ', d2),\n                  getW(wF, wR, wC, ' +
        d +
        ' + 1, d2),\n                  getW(wF, wR, wC, ' +
        d +
        ' + 2, d2)\n                );\n                dotProd += dot(xValues, wValues);\n              }\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Za = function (t, e, n, r) {
      void 0 === e && (e = !1),
        void 0 === n && (n = null),
        void 0 === r && (r = !1),
        (this.variableNames = ['x', 'W']),
        (this.outputShape = t.outShape)
      var o = t.inHeight,
        a = t.inWidth,
        i = t.padInfo.top,
        s = t.padInfo.left,
        u = t.strideHeight,
        c = t.strideWidth,
        l = t.dilationHeight,
        h = t.dilationWidth,
        f = t.filterHeight,
        d = t.filterWidth,
        p = t.outChannels / t.inChannels,
        v = '',
        m = ''
      n &&
        ((v = r
          ? 'float activation(float a) {\n          float b = getPreluActivationWeightsAtOutCoords();\n          ' +
            n +
            '\n        }'
          : '\n          float activation(float x) {\n            ' +
            n +
            '\n          }\n        '),
        (m = 'result = activation(result);'))
      var g = e ? 'result += getBiasAtOutCoords();' : ''
      e && this.variableNames.push('bias'),
        r && this.variableNames.push('preluActivationWeights'),
        (this.userCode =
          '\n      ' +
          v +
          '\n\n      const ivec2 strides = ivec2(' +
          u +
          ', ' +
          c +
          ');\n      const ivec2 pads = ivec2(' +
          i +
          ', ' +
          s +
          ');\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int batch = coords.x;\n        ivec2 xRCCorner = coords.yz * strides - pads;\n        int d2 = coords.w;\n        int d1 = d2 / ' +
          p +
          ';\n        int q = d2 - d1 * ' +
          p +
          ';\n\n        int xRCorner = xRCCorner.x;\n        int xCCorner = xRCCorner.y;\n\n        // Convolve x(?, ?, d1) with w(:, :, d1, q) to get y(yR, yC, d2).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n        // TO DO(dsmilkov): Flatten the two for loops and vec4 the operations.\n        for (int wR = 0; wR < ' +
          f +
          '; wR++) {\n          int xR = xRCorner + wR * ' +
          l +
          ';\n\n          if (xR < 0 || xR >= ' +
          o +
          ') {\n            continue;\n          }\n\n          for (int wC = 0; wC < ' +
          d +
          '; wC++) {\n            int xC = xCCorner + wC * ' +
          h +
          ';\n\n            if (xC < 0 || xC >= ' +
          a +
          ') {\n              continue;\n            }\n\n            float xVal = getX(batch, xR, xC, d1);\n            float wVal = getW(wR, wC, d1, q);\n            dotProd += xVal * wVal;\n          }\n        }\n\n        float result = dotProd;\n        ' +
          g +
          '\n        ' +
          m +
          '\n        setOutput(result);\n      }\n    ')
    },
    ti = function (t, e, n, r) {
      void 0 === e && (e = !1),
        void 0 === n && (n = null),
        void 0 === r && (r = !1),
        (this.variableNames = ['x', 'W']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = t.outShape)
      for (
        var o = t.inHeight,
          a = t.inWidth,
          i = t.padInfo.top,
          s = t.padInfo.left,
          u = t.strideHeight,
          c = t.strideWidth,
          l = t.dilationHeight,
          h = t.dilationWidth,
          f = t.filterHeight,
          d = t.filterWidth,
          p = d,
          v = 'int xR; int xC; int xCOffset;',
          m = 0;
        m < f;
        m++
      )
        for (var g = 0; g < d; g++)
          v +=
            '\n          vec4 xTexelR' +
            m +
            'C' +
            2 * g +
            ' = vec4(0.);\n          vec4 wR' +
            m +
            'C' +
            g +
            ' = vec4(0.);\n          vec4 xR' +
            m +
            'C' +
            g +
            ' = vec4(0.);'
      for (m = 0; m < f; m++)
        for (var y = 0; y < p; y++) {
          if (
            ((v +=
              '\n          xR = xRCorner + ' +
              m * l +
              ';\n          xC = xCCorner + ' +
              (g = 2 * y) * h +
              ';\n        '),
            1 === c)
          ) {
            if (
              g < d &&
              ((v +=
                s % 2 == 1
                  ? '\n                xCOffset = xC + 1;\n                if(xR >= 0 && xR < ' +
                    o +
                    ' && xCOffset >= 0 && xCOffset < ' +
                    a +
                    ') {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    " = getX(batch, xR, xCOffset, d1);\n\n                  // Need to manually clear unused channels in case\n                  // we're reading from recycled texture.\n                  if(xCOffset + 1 >= " +
                    a +
                    ') {\n                    xTexelR' +
                    m +
                    'C' +
                    g +
                    '.zw = vec2(0.);\n                  }\n                } else {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(0.);\n                }\n\n                xCOffset = xC + 1 - 2;\n                if(xR >= 0 && xR < ' +
                    o +
                    ' && xCOffset >= 0 && xCOffset < ' +
                    a +
                    ") {\n                  vec4 previous = getX(batch, xR, xCOffset, d1);\n\n                  // Need to manually clear unused channels in case\n                  // we're reading from recycled texture.\n                  if(xCOffset + 1 >= " +
                    a +
                    ') {\n                    previous.zw = vec2(0.);\n                  }\n\n                  xR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(previous.zw, xTexelR' +
                    m +
                    'C' +
                    g +
                    '.xy);\n                } else {\n                  xR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(0, 0, xTexelR' +
                    m +
                    'C' +
                    g +
                    '.xy);\n                }\n              '
                  : '\n                if(xR >= 0 && xR < ' +
                    o +
                    ' && xC >= 0 && xC < ' +
                    a +
                    ') {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = getX(batch, xR, xC, d1);\n                } else {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(0.);\n                }\n\n                xR' +
                    m +
                    'C' +
                    g +
                    ' = xTexelR' +
                    m +
                    'C' +
                    g +
                    ';\n              '),
              g + 1 < d)
            ) {
              var x = s % 2 == 0 ? b(h) : h
              ;(h % 2 == 0 && s % 2 == 1) || (h % 2 != 0 && s % 2 != 1)
                ? ((v +=
                    '\n                  xCOffset = xC + ' +
                    (s % 2) +
                    ' + ' +
                    x +
                    ';\n\n                  if(xR >= 0 && xR < ' +
                    o +
                    ' &&\n                    xCOffset >= 0 && xCOffset < ' +
                    a +
                    ') {\n                    xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ' = getX(batch, xR, xCOffset, d1);\n                  }\n                '),
                  h > 1 &&
                    (v +=
                      '\n                    xCOffset -= 2;\n                    if(xR >= 0 && xR < ' +
                      o +
                      ' &&\n                      xCOffset >= 0 && xCOffset < ' +
                      a +
                      ') {\n                      xTexelR' +
                      m +
                      'C' +
                      g +
                      ' = getX(batch, xR, xCOffset, d1);\n                    } else {\n                      xTexelR' +
                      m +
                      'C' +
                      g +
                      ' = vec4(0.);\n                    }\n                  '),
                  (v +=
                    '\n                  xR' +
                    m +
                    'C' +
                    (g + 1) +
                    ' = vec4(\n                    xTexelR' +
                    m +
                    'C' +
                    g +
                    '.zw, xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    '.xy);\n                '))
                : (v +=
                    '\n                  xCOffset = xC + ' +
                    x +
                    ';\n\n                  if(xR >= 0 && xR < ' +
                    o +
                    ' &&\n                    xCOffset >= 0 && xCOffset < ' +
                    a +
                    ') {\n                    xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ' = getX(batch, xR, xCOffset, d1);\n                  }\n\n                  xR' +
                    m +
                    'C' +
                    (g + 1) +
                    ' = xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ';\n                ')
            }
          } else
            g < d &&
              ((v +=
                '\n              if(xR >= 0 && xR < ' +
                o +
                ') {\n            '),
              s % 2 == 1
                ? ((v +=
                    '\n                xCOffset = xC + 1 - ' +
                    c +
                    ';\n                if(xCOffset >= 0 && xCOffset < ' +
                    a +
                    ') {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = getX(batch, xR, xCOffset, d1);\n                } else {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(0.);\n                }\n\n                if(xC + 1 >= 0 && xC + 1 < ' +
                    a +
                    ') {\n                  xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ' = getX(batch, xR, xC + 1, d1);\n                } else {\n                  xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ' = vec4(0.);\n                }\n\n                xR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    '.zw, xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    '.zw);\n              '),
                  g + 1 < d &&
                    (v +=
                      '\n                  vec4 final = vec4(0.);\n                  xCOffset = xC + 1 + ' +
                      c +
                      ';\n                  if(xCOffset >= 0 && xCOffset < ' +
                      a +
                      ') {\n                    final = getX(batch, xR, xCOffset, d1);\n                  }\n                  xR' +
                      m +
                      'C' +
                      (g + 1) +
                      ' = vec4(xTexelR' +
                      m +
                      'C' +
                      (g + 2) +
                      '.xy, final.xy);\n                '))
                : ((v +=
                    '\n                if(xC >= 0 && xC < ' +
                    a +
                    ') {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = getX(batch, xR, xC, d1);\n                } else {\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(0.);\n                }\n\n                xCOffset = xC + ' +
                    c +
                    ';\n                if(xCOffset >= 0 && xCOffset < ' +
                    a +
                    ') {\n                  xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ' = getX(batch, xR, xCOffset, d1);\n                } else {\n                  xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    ' = vec4(0.);\n                }\n\n                xR' +
                    m +
                    'C' +
                    g +
                    ' = vec4(\n                  xTexelR' +
                    m +
                    'C' +
                    g +
                    '.xy, xTexelR' +
                    m +
                    'C' +
                    (g + 2) +
                    '.xy);\n              '),
                  g + 1 < d &&
                    (v +=
                      '\n                  xR' +
                      m +
                      'C' +
                      (g + 1) +
                      ' = vec4(\n                    xTexelR' +
                      m +
                      'C' +
                      g +
                      '.zw, xTexelR' +
                      m +
                      'C' +
                      (g + 2) +
                      '.zw);\n                ')),
              (v += '}'))
          g < d &&
            ((v +=
              '\n            vec4 wTexelR' +
              m +
              'C' +
              g +
              ' = getW(' +
              m +
              ', ' +
              g +
              ', d1, q);\n            wR' +
              m +
              'C' +
              g +
              ' = vec4(wTexelR' +
              m +
              'C' +
              g +
              '.xz, wTexelR' +
              m +
              'C' +
              g +
              '.xz);\n          '),
            g + 1 < d &&
              (v +=
                '\n              vec4 wTexelR' +
                m +
                'C' +
                (g + 1) +
                ' = getW(' +
                m +
                ', ' +
                (g + 1) +
                ', d1, q);\n              wR' +
                m +
                'C' +
                (g + 1) +
                ' =\n                vec4(wTexelR' +
                m +
                'C' +
                (g + 1) +
                '.xz, wTexelR' +
                m +
                'C' +
                (g + 1) +
                '.xz);'))
        }
      for (m = 0; m < f; m++)
        for (g = 0; g < d; g++)
          v += 'dotProd += xR' + m + 'C' + g + ' * wR' + m + 'C' + g + ';'
      var w = '',
        C = ''
      n &&
        ((w = r
          ? 'vec4 activation(vec4 a) {\n          vec4 b = getPreluActivationWeightsAtOutCoords();\n          ' +
            n +
            '\n        }'
          : 'vec4 activation(vec4 x) {\n          ' + n + '\n        }'),
        (C = 'result = activation(result);'))
      var E = e ? 'result += getBiasAtOutCoords();' : ''
      e && this.variableNames.push('bias'),
        r && this.variableNames.push('preluActivationWeights'),
        (this.userCode =
          '\n      ' +
          w +
          '\n\n      const ivec2 strides = ivec2(' +
          u +
          ', ' +
          c +
          ');\n      const ivec2 pads = ivec2(' +
          i +
          ', ' +
          s +
          ');\n\n      void main() {\n\n        ivec4 coords = getOutputCoords();\n        int batch = coords.x;\n        ivec2 xRCCorner = coords.yz * strides - pads;\n        int d2 = coords.w;\n        int d1 = d2;\n        int q = 0;\n        int xRCorner = xRCCorner.x;\n        int xCCorner = xRCCorner.y;\n\n        vec4 dotProd = vec4(0.);\n\n        ' +
          v +
          '\n\n        vec4 result = dotProd;\n        ' +
          E +
          '\n        ' +
          C +
          '\n        setOutput(result);\n      }\n    ')
    },
    ei = function (t, e, n, r, o) {
      ;(this.variableNames = ['Image', 'Boxes', 'BoxInd']),
        (this.outputShape = [])
      var a = t[0],
        i = t[1],
        s = t[2],
        u = t[3],
        c = e[0],
        l = n[0],
        h = n[1]
      this.outputShape = [c, l, h, u]
      var f = 'bilinear' === r ? 1 : 0,
        d = [i - 1 + '.0', s - 1 + '.0'],
        p = d[0],
        v = d[1],
        m =
          l > 1
            ? [
                '' + (i - 1) / (l - 1),
                '(y2-y1) * height_ratio',
                'y1*' + p + ' + float(y)*(height_scale)',
              ]
            : ['0.0', '0.0', '0.5 * (y1+y2) * ' + p],
        g = m[0],
        y = m[1],
        x = m[2],
        b =
          h > 1
            ? [
                '' + (s - 1) / (h - 1),
                '(x2-x1) * width_ratio',
                'x1*' + v + ' + float(x)*(width_scale)',
              ]
            : ['0.0', '0.0', '0.5 * (x1+x2) * ' + v],
        w = b[0],
        C = b[1],
        E = b[2]
      this.userCode =
        '\n      const float height_ratio = float(' +
        g +
        ');\n      const float width_ratio = float(' +
        w +
        ');\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int y = coords[1];\n        int x = coords[2];\n        int d = coords[3];\n\n        // get box vals\n        float y1 = getBoxes(b,0);\n        float x1 = getBoxes(b,1);\n        float y2 = getBoxes(b,2);\n        float x2 = getBoxes(b,3);\n\n        // get image in batch index\n        int bInd = round(getBoxInd(b));\n        if(bInd < 0 || bInd >= ' +
        a +
        ') {\n          return;\n        }\n\n        float height_scale = ' +
        y +
        ';\n        float width_scale = ' +
        C +
        ';\n\n        float in_y = ' +
        x +
        ';\n        if( in_y < 0.0 || in_y > ' +
        p +
        ' ) {\n          setOutput(float(' +
        o +
        '));\n          return;\n        }\n        float in_x = ' +
        E +
        ';\n        if( in_x < 0.0 || in_x > ' +
        v +
        ' ) {\n          setOutput(float(' +
        o +
        '));\n          return;\n        }\n\n        vec2 sourceFracIndexCR = vec2(in_x,in_y);\n        if(' +
        f +
        ' == 1) {\n          // Compute the four integer indices.\n          ivec2 sourceFloorCR = ivec2(sourceFracIndexCR);\n          ivec2 sourceCeilCR = ivec2(ceil(sourceFracIndexCR));\n\n          float topLeft = getImage(b, sourceFloorCR.y, sourceFloorCR.x, d);\n          float bottomLeft = getImage(b, sourceCeilCR.y, sourceFloorCR.x, d);\n          float topRight = getImage(b, sourceFloorCR.y, sourceCeilCR.x, d);\n          float bottomRight = getImage(b, sourceCeilCR.y, sourceCeilCR.x, d);\n\n          vec2 fracCR = sourceFracIndexCR - vec2(sourceFloorCR);\n\n          float top = topLeft + (topRight - topLeft) * fracCR.x;\n          float bottom = bottomLeft + (bottomRight - bottomLeft) * fracCR.x;\n          float newValue = top + (bottom - top) * fracCR.y;\n          setOutput(newValue);\n        } else {\n          // Compute the coordinators of nearest neighbor point.\n          ivec2 sourceNearestCR = ivec2(floor(\n            sourceFracIndexCR + vec2(0.5,0.5)));\n          float newValue = getImage(b, sourceNearestCR.y, sourceNearestCR.x, d);\n          setOutput(newValue);\n        }\n      }\n    '
    },
    ni = function (t, e, n) {
      ;(this.variableNames = ['x']), (this.outputShape = t)
      var r = t.length,
        o = t[t.length - 1],
        a = n ? '<' : '>'
      this.userCode =
        '\n      int getIndex(int i) {\n        ' +
        (n ? 'return ' + o + ' -i - 1;' : 'return i;') +
        '\n      }\n\n      void main() {\n        ' +
        wa(r) +
        ' coords = getOutputCoords();\n        int end = ' +
        ri(r, 'coords') +
        ';\n        float val = 0.0;\n        for (int i = ' +
        o +
        ' - 1; i >= 0; i -= 1) {\n          int idx = getIndex(i);\n          if (idx ' +
        a +
        ' end) {\n            continue;\n          }\n          if (idx == end && ' +
        e +
        ') {\n            continue;\n          }\n          ' +
        ri(r, 'coords') +
        ' = idx;\n          val += getX(' +
        (function (t, e) {
          if (1 === t) return '' + e
          if (2 === t) return e + '.x, ' + e + '.y'
          if (3 === t) return e + '.x, ' + e + '.y, ' + e + '.z'
          if (4 === t) return e + '.x, ' + e + '.y, ' + e + '.z, ' + e + '.w'
          throw Error('Cumulative sum for rank ' + t + ' is not yet supported')
        })(r, 'coords') +
        ');\n        }\n        setOutput(val);\n      }\n    '
    }
  function ri(t, e) {
    if (1 === t) return '' + e
    if (2 === t) return e + '.y'
    if (3 === t) return e + '.z'
    if (4 === t) return e + '.w'
    throw Error('Cumulative sum for rank ' + t + ' is not yet supported')
  }
  var oi = function (t) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !1),
        (this.packedOutput = !0),
        (this.outPackingScheme = Vt.DENSE)
      var e = Yt(t),
        n = ua()
      ;(this.outputShape = t),
        (this.userCode =
          '\n      ivec3 outCoordsFromFlatIndex(int index) {\n        ' +
          ca(['r', 'c', 'd'], t) +
          '\n        return ivec3(r, c, d);\n      }\n\n      void main() {\n        ivec2 resTexRC = ivec2(resultUV.yx *\n          vec2(' +
          e[0] +
          ', ' +
          e[1] +
          '));\n        int index = 4 * (resTexRC.x * ' +
          e[1] +
          ' + resTexRC.y);\n\n        vec4 result = vec4(0.);\n\n        for (int i=0; i<4; i++) {\n          int flatIndex = index + i;\n          ivec3 rc = outCoordsFromFlatIndex(flatIndex);\n          result[i] = getA(rc.x, rc.y, rc.z);\n        }\n\n        ' +
          n.output +
          ' = result;\n      }\n    ')
    },
    ai = function (t) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outPackingScheme = Vt.DENSE)
      var e = Yt(t),
        n = ua()
      ;(this.outputShape = t),
        (this.userCode =
          '\n      ivec3 outCoordsFromFlatIndex(int index) {\n        ' +
          ca(['r', 'c', 'd'], t) +
          '\n        return ivec3(r, c, d);\n      }\n\n      void main() {\n        ivec2 resTexRC = ivec2(resultUV.yx *\n          vec2(' +
          e[0] +
          ', ' +
          e[1] +
          '));\n        int index = 4 * (resTexRC.x * ' +
          e[1] +
          ' + resTexRC.y);\n\n        vec4 result = vec4(0.);\n\n        for (int i=0; i<4; i++) {\n          int flatIndex = index + i;\n          ivec3 rc = outCoordsFromFlatIndex(flatIndex);\n          result[i] = getChannel(getA(rc.x, rc.y, rc.z), vec2(rc.y, rc.z));\n        }\n\n        ' +
          n.output +
          ' = result;\n      }\n    ')
    },
    ii = (function () {
      function t(t, e, n) {
        ;(this.variableNames = ['x']),
          (this.outputShape = []),
          (this.outputShape = t),
          (this.blockSize = e),
          (this.dataFormat = n),
          (this.userCode =
            '\n    void main() {\n      ivec4 coords = getOutputCoords();\n      int b = coords[0];\n      int h = ' +
            this.getHeightCoordString() +
            ';\n      int w = ' +
            this.getWidthCoordString() +
            ';\n      int d = ' +
            this.getDepthCoordString() +
            ';\n\n      int in_h = h / ' +
            e +
            ';\n      int offset_h = imod(h, ' +
            e +
            ');\n      int in_w = w / ' +
            e +
            ';\n      int offset_w = imod(w, ' +
            e +
            ');\n      int offset_d = (offset_h * ' +
            e +
            ' + offset_w) *\n        ' +
            this.getOutputDepthSize() +
            ';\n      int in_d = d + offset_d;\n\n      float result = ' +
            this.getInputSamplingString() +
            ';\n      setOutput(result);\n    }\n  ')
      }
      return (
        (t.prototype.getHeightCoordString = function () {
          return 'NHWC' === this.dataFormat ? 'coords[1]' : 'coords[2]'
        }),
        (t.prototype.getWidthCoordString = function () {
          return 'NHWC' === this.dataFormat ? 'coords[2]' : 'coords[3]'
        }),
        (t.prototype.getDepthCoordString = function () {
          return 'NHWC' === this.dataFormat ? 'coords[3]' : 'coords[1]'
        }),
        (t.prototype.getOutputDepthSize = function () {
          return 'NHWC' === this.dataFormat
            ? this.outputShape[3]
            : this.outputShape[1]
        }),
        (t.prototype.getInputSamplingString = function () {
          return 'NHWC' === this.dataFormat
            ? 'getX(b, in_h, in_w, in_d)'
            : 'getX(b, in_d, in_h, in_w)'
        }),
        t
      )
    })(),
    si = function (t) {
      ;(this.variableNames = ['X']),
        (this.outputShape = [t, t]),
        (this.userCode =
          '\n      void main() {\n          ivec2 coords = getOutputCoords();\n          float val = coords[0] == coords[1] ? getX(coords[0]) : 0.0;\n          setOutput(val);\n      }\n    ')
    },
    ui = function (t) {
      ;(this.variableNames = ['A']), (this.outTexUsage = zt.DOWNLOAD)
      var e = ua()
      ;(this.outputShape = t),
        (this.userCode =
          '\n      ' +
          ha +
          '\n\n      void main() {\n        float x = getAAtOutCoords();\n        ' +
          e.output +
          ' = encode_float(x);\n      }\n    ')
    },
    ci = function (t) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !1),
        (this.outTexUsage = zt.DOWNLOAD)
      var e = ua()
      ;(this.outputShape = t),
        (this.userCode =
          '\n      ' +
          ha +
          '\n\n      void main() {\n        ivec3 coords = getOutputCoords();\n        float x = getChannel(getAAtOutCoords(), vec2(coords.y, coords.z));\n        ' +
          e.output +
          ' = encode_float(x);\n      }\n    ')
    },
    li = function (t, e, n) {
      void 0 === n && (n = !1), (this.variableNames = ['A'])
      var r = ua(),
        o = e[0],
        a = e[1]
      this.outputShape = t
      var i = 'result'
      n && (i = 'floor(result * 255. + 0.5)'),
        (this.userCode =
          '\n      ' +
          la(t) +
          '\n\n      void main() {\n        ivec3 coords = getOutputCoords();\n\n        int flatIndex = getFlatIndex(coords);\n        int offset = imod(flatIndex, 4);\n\n        flatIndex = idiv(flatIndex, 4, 1.);\n        \n        int r = flatIndex / ' +
          a +
          ';\n        int c = imod(flatIndex, ' +
          a +
          ');\n        vec2 uv = (vec2(c, r) + halfCR) / vec2(' +
          a +
          '.0, ' +
          o +
          '.0);\n        vec4 values = ' +
          r.texture2D +
          '(A, uv);\n\n        float result;\n\n        if(offset == 0) {\n          result = values[0];\n        } else if(offset == 1) {\n          result = values[1];\n        } else if(offset == 2) {\n          result = values[2];\n        } else {\n          result = values[3];\n        }\n\n        ' +
          r.output +
          ' = vec4(' +
          i +
          ', 0., 0., 0.);\n      }\n    ')
    },
    hi = function (t, e, n) {
      void 0 === n && (n = !1),
        (this.variableNames = ['A']),
        (this.packedInputs = !1),
        (this.packedOutput = !0)
      var r = ua(),
        o = e[0],
        a = e[1]
      this.outputShape = t
      var i = '',
        s = 'result'
      n && (s = 'floor(result * 255. + 0.5)')
      for (var u = 0; u <= 1; u++)
        for (var c = 0; c <= 1; c++) {
          var l = 2 * u + c
          i +=
            '\n          localCoords = coords;\n          if(localCoords[2] + ' +
            c +
            ' < ' +
            t[2] +
            ') {\n            localCoords[2] += ' +
            c +
            ';\n            if(localCoords[1] + ' +
            u +
            ' < ' +
            t[1] +
            ') {\n              localCoords[1] += ' +
            u +
            ';\n\n              flatIndex = getFlatIndex(localCoords);\n              offset = imod(flatIndex, 4);\n\n              flatIndex = idiv(flatIndex, 4, 1.);\n\n              r = flatIndex / ' +
            a +
            ';\n              c = imod(flatIndex, ' +
            a +
            ');\n              uv = (vec2(c, r) + halfCR) / vec2(' +
            a +
            '.0, ' +
            o +
            '.0);\n              values = ' +
            r.texture2D +
            '(A, uv);\n\n              if(offset == 0) {\n                result[' +
            l +
            '] = values[0];\n              } else if(offset == 1) {\n                result[' +
            l +
            '] = values[1];\n              } else if(offset == 2) {\n                result[' +
            l +
            '] = values[2];\n              } else {\n                result[' +
            l +
            '] = values[3];\n              }\n            }\n          }\n        '
        }
      this.userCode =
        '\n      ' +
        la(t) +
        '\n\n      void main() {\n        ivec3 coords = getOutputCoords();\n\n        vec4 result = vec4(0.);\n        int flatIndex, r, c, offset;\n        ivec3 localCoords;\n        vec2 uv;\n        vec4 values;\n\n        ' +
        i +
        '\n\n        ' +
        r.output +
        ' = ' +
        s +
        ';\n      }\n    '
    },
    fi = 'return real * expR - imag * expI;',
    di = 'return real * expI + imag * expR;',
    pi = function (t, e, n) {
      this.variableNames = ['real', 'imag']
      var r = e[1]
      this.outputShape = e
      var o = n ? '2.0 * ' + Math.PI : '-2.0 * ' + Math.PI,
        a = n ? r + '.0' : '1.0'
      this.userCode =
        '\n      const float exponentMultiplier = ' +
        o +
        ';\n\n      float unaryOpComplex(float real, float expR, float imag, float expI) {\n        ' +
        t +
        '\n      }\n\n      float mulMatDFT(int batch, int index) {\n        float indexRatio = float(index) / float(' +
        r +
        ');\n        float exponentMultiplierTimesIndexRatio =\n            exponentMultiplier * indexRatio;\n\n        float result = 0.0;\n\n        for (int i = 0; i < ' +
        r +
        '; i++) {\n          // x = (-2|2 * PI / N) * index * i;\n          float x = exponentMultiplierTimesIndexRatio * float(i);\n          float expR = cos(x);\n          float expI = sin(x);\n          float real = getReal(batch, i);\n          float imag = getImag(batch, i);\n\n          result +=\n              unaryOpComplex(real, expR, imag, expI) / ' +
        a +
        ';\n        }\n\n        return result;\n      }\n\n      void main() {\n        ivec2 coords = getOutputCoords();\n        setOutput(mulMatDFT(coords[0], coords[1]));\n      }\n    '
    },
    vi = (function () {
      function t(t, e) {
        ;(this.outputShape = []),
          (this.variableNames = ['x']),
          (this.outputShape = t),
          (this.userCode =
            '\n      uniform float value;\n      void main() {\n        // Input can be obtained from uniform value.\n        setOutput(value);\n      }\n    ')
      }
      return (
        (t.prototype.getCustomSetupFunc = function (t) {
          var e = this
          return function (n, r) {
            null == e.valueLoc &&
              (e.valueLoc = n.getUniformLocationNoThrow(r, 'value')),
              n.gl.uniform1f(e.valueLoc, t)
          }
        }),
        t
      )
    })(),
    mi = function (t, e, n) {
      this.variableNames = ['A', 'indices']
      var r = t.slice()
      ;(r[n] = e), (this.outputShape = r), (this.rank = r.length)
      var o = wa(this.rank),
        a = (function (t, e) {
          var n = t.length
          if (n > 4)
            throw Error('Gather for rank ' + n + ' is not yet supported')
          if (1 === n) return 'int(getIndices(resRC))'
          for (
            var r = ['resRC.x', 'resRC.y', 'resRC.z', 'resRC.w'], o = [], a = 0;
            a < t.length;
            a++
          )
            a === e
              ? o.push('int(getIndices(' + r[a] + '))')
              : o.push('' + r[a])
          return o.join()
        })(t, n)
      this.userCode =
        '\n      void main() {\n        ' +
        o +
        ' resRC = getOutputCoords();\n        setOutput(getA(' +
        a +
        '));\n      }\n    '
    }
  var gi = function (t, e, n) {
    ;(this.sliceDim = t),
      (this.strides = e),
      (this.variableNames = ['x', 'indices']),
      (this.outputShape = n)
    var r = wa(e.length),
      o = wa(n.length),
      a = this.sliceDim > 1 ? 'strides[j]' : 'strides'
    this.userCode =
      '\n        ' +
      r +
      ' strides = ' +
      r +
      '(' +
      this.strides +
      ');\n         void main() {\n          ' +
      o +
      ' coords = getOutputCoords();\n          int flattenIndex = 0;\n          for (int j = 0; j < ' +
      this.sliceDim +
      '; j++) {\n            int index = round(getIndices(coords[0], j));\n            flattenIndex += index * ' +
      a +
      ';\n          }\n          setOutput(getX(flattenIndex, coords[1]));\n        }\n      '
  }
  function yi(t, e) {
    var n = ua()
    return oe(
      t,
      e,
      n.version +
        '\n    precision highp float;\n    ' +
        n.attribute +
        ' vec3 clipSpacePos;\n    ' +
        n.attribute +
        ' vec2 uv;\n    ' +
        n.varyingVs +
        ' vec2 resultUV;\n\n    void main() {\n      gl_Position = vec4(clipSpacePos, 1);\n      resultUV = uv;\n    }'
    )
  }
  function xi(t, e) {
    return fe(
      t,
      e,
      new Float32Array([
        -1, 1, 0, 0, 1, -1, -1, 0, 0, 0, 1, 1, 0, 1, 1, 1, -1, 0, 1, 0,
      ])
    )
  }
  function bi(t, e) {
    return de(t, e, new Uint16Array([0, 1, 2, 2, 1, 3]))
  }
  function wi(t, e, n, r, o, a, i) {
    ve(n, r)
    var s = pe(t, e),
      u = t.TEXTURE_2D
    return (
      Jt(t, e, function () {
        return t.bindTexture(u, s)
      }),
      Jt(t, e, function () {
        return t.texParameteri(u, t.TEXTURE_WRAP_S, t.CLAMP_TO_EDGE)
      }),
      Jt(t, e, function () {
        return t.texParameteri(u, t.TEXTURE_WRAP_T, t.CLAMP_TO_EDGE)
      }),
      Jt(t, e, function () {
        return t.texParameteri(u, t.TEXTURE_MIN_FILTER, t.NEAREST)
      }),
      Jt(t, e, function () {
        return t.texParameteri(u, t.TEXTURE_MAG_FILTER, t.NEAREST)
      }),
      Jt(t, e, function () {
        return t.texImage2D(u, 0, o, n, r, 0, a, i, null)
      }),
      Jt(t, e, function () {
        return t.bindTexture(t.TEXTURE_2D, null)
      }),
      s
    )
  }
  function Ci(t, e, n, r, o) {
    var a = Xt(n, r)
    return wi(
      t,
      e,
      a[0],
      a[1],
      o.internalFormatFloat,
      o.textureFormatFloat,
      t.FLOAT
    )
  }
  function Ei(t, e, n, r, o) {
    var a = Xt(n, r)
    return wi(
      t,
      e,
      a[0],
      a[1],
      o.internalFormatHalfFloat,
      o.textureFormatFloat,
      o.textureTypeHalfFloat
    )
  }
  function Ri(t, e, n, r, o) {
    var a = Xt(n, r)
    return wi(t, e, a[0], a[1], t.RGBA, t.RGBA, t.UNSIGNED_BYTE)
  }
  function Ii(t, e, n, r, o) {
    var a = $t(n, r)
    return wi(t, e, a[0], a[1], o.internalFormatPackedFloat, t.RGBA, t.FLOAT)
  }
  function ki(t, e, n, r, o) {
    var a = $t(n, r)
    return wi(
      t,
      e,
      a[0],
      a[1],
      o.internalFormatPackedHalfFloat,
      t.RGBA,
      o.textureTypeHalfFloat
    )
  }
  function Si(t, e, n, r) {
    return (
      Jt(t, e, function () {
        return t.bindBuffer(t.ARRAY_BUFFER, r)
      }),
      ge(t, e, n, 'clipSpacePos', r, 3, 20, 0) &&
        ge(t, e, n, 'uv', r, 2, 20, 12)
    )
  }
  function Ai(t, e, n, r, o, a, i) {
    var s, u, c
    Jt(t, e, function () {
      return t.bindTexture(t.TEXTURE_2D, n)
    }),
      a instanceof Uint8Array
        ? ((s = new Uint8Array(r * o * 4)), (u = t.UNSIGNED_BYTE), (c = t.RGBA))
        : ((s = new Float32Array(r * o * 4)),
          (u = t.FLOAT),
          (c = i.internalFormatPackedFloat)),
      s.set(a),
      Jt(t, e, function () {
        return t.texImage2D(t.TEXTURE_2D, 0, c, r, o, 0, t.RGBA, u, s)
      }),
      Jt(t, e, function () {
        return t.bindTexture(t.TEXTURE_2D, null)
      })
  }
  function Di(t, e, n, r) {
    Jt(t, e, function () {
      return t.bindTexture(t.TEXTURE_2D, n)
    }),
      r.data instanceof Uint8Array
        ? Jt(t, e, function () {
            return t.texImage2D(
              t.TEXTURE_2D,
              0,
              t.RGBA,
              r.width,
              r.height,
              0,
              t.RGBA,
              t.UNSIGNED_BYTE,
              r.data
            )
          })
        : Jt(t, e, function () {
            return t.texImage2D(
              t.TEXTURE_2D,
              0,
              t.RGBA,
              t.RGBA,
              t.UNSIGNED_BYTE,
              r
            )
          }),
      Jt(t, e, function () {
        return t.bindTexture(t.TEXTURE_2D, null)
      })
  }
  function Ti(t, e, n, r, o) {
    var a = t.createBuffer()
    Jt(t, e, function () {
      return t.bindBuffer(t.PIXEL_PACK_BUFFER, a)
    })
    var i = 16 * n * r
    return (
      Jt(t, e, function () {
        return t.bufferData(t.PIXEL_PACK_BUFFER, i, t.STREAM_READ)
      }),
      Jt(t, e, function () {
        return t.readPixels(0, 0, r, n, t.RGBA, t.FLOAT, 0)
      }),
      Jt(t, e, function () {
        return t.bindBuffer(t.PIXEL_PACK_BUFFER, null)
      }),
      a
    )
  }
  function Ni(t, e, n) {
    var r = t,
      o = new Float32Array(n)
    return (
      r.bindBuffer(r.PIXEL_PACK_BUFFER, e),
      r.getBufferSubData(r.PIXEL_PACK_BUFFER, 0, o),
      r.bindBuffer(r.PIXEL_PACK_BUFFER, null),
      o
    )
  }
  function Fi(t, e, n, r, o) {
    var a = Xt(n, r),
      i = a[0],
      s = a[1],
      u = new Uint8Array(n * r * 4)
    return (
      Jt(t, e, function () {
        return t.readPixels(
          0,
          0,
          i,
          s,
          o.downloadTextureFormat,
          t.UNSIGNED_BYTE,
          u
        )
      }),
      new Float32Array(u.buffer)
    )
  }
  function _i(t, e, n, r, o, a, i, s) {
    var u = t,
      c = new Float32Array(
        (function (t, e) {
          var n = $t(t, e)
          return n[0] * n[1] * 4
        })(a, i)
      )
    return (
      u.bindBuffer(u.PIXEL_PACK_BUFFER, e),
      u.getBufferSubData(u.PIXEL_PACK_BUFFER, 0, c),
      u.bindBuffer(u.PIXEL_PACK_BUFFER, null),
      c
    )
  }
  function Oi(t, e, n, r) {
    var o = new Float32Array(n * r * 4)
    return (
      Jt(t, e, function () {
        return t.readPixels(0, 0, r, n, t.RGBA, t.FLOAT, o)
      }),
      o
    )
  }
  var Mi = Object.freeze({
      createVertexShader: yi,
      createVertexBuffer: xi,
      createIndexBuffer: bi,
      createFloat32MatrixTexture: Ci,
      createFloat16MatrixTexture: Ei,
      createUnsignedBytesMatrixTexture: Ri,
      createPackedMatrixTexture: Ii,
      createFloat16PackedMatrixTexture: ki,
      bindVertexProgramAttributeStreams: Si,
      uploadDenseMatrixToTexture: Ai,
      uploadPixelDataToTexture: Di,
      createBufferFromOutputTexture: Ti,
      downloadFloat32MatrixFromBuffer: Ni,
      downloadByteEncodedFloatMatrixFromOutputTexture: Fi,
      downloadPackedMatrixFromBuffer: _i,
      downloadMatrixFromPackedOutputTexture: Oi,
    }),
    Bi = (function () {
      function t(t) {
        ;(this.outputTexture = null),
          (this.program = null),
          (this.disposed = !1),
          (this.vertexAttrsAreBound = !1),
          (this.itemsToPoll = [])
        var e = i().getNumber('WEBGL_VERSION')
        null != t ? ((this.gl = t), Kt(e, t)) : (this.gl = jt(e))
        var n = 'WEBGL_color_buffer_float'
        if (1 === i().getNumber('WEBGL_VERSION')) {
          if (
            ((this.textureFloatExtension = re(
              this.gl,
              this.debug,
              'OES_texture_float'
            )),
            Pe(this.gl, 'OES_texture_half_float'))
          )
            this.textureHalfFloatExtension = re(
              this.gl,
              this.debug,
              'OES_texture_half_float'
            )
          else if (i().get('WEBGL_FORCE_F16_TEXTURES'))
            throw new Error(
              'GL context does not support half float textures, yet the environment flag WEBGL_FORCE_F16_TEXTURES is set to true.'
            )
          if (
            ((this.colorBufferFloatExtension = this.gl.getExtension(n)),
            Pe(this.gl, 'EXT_color_buffer_half_float'))
          )
            this.colorBufferHalfFloatExtension = re(
              this.gl,
              this.debug,
              'EXT_color_buffer_half_float'
            )
          else if (i().get('WEBGL_FORCE_F16_TEXTURES'))
            throw new Error(
              'GL context does not support color renderable half floats, yet the environment flag WEBGL_FORCE_F16_TEXTURES is set to true.'
            )
        } else if (((n = 'EXT_color_buffer_float'), Pe(this.gl, n)))
          this.colorBufferFloatExtension = this.gl.getExtension(n)
        else {
          if (!Pe(this.gl, 'EXT_color_buffer_half_float'))
            throw new Error(
              'GL context does not support color renderable floats'
            )
          this.colorBufferHalfFloatExtension = this.gl.getExtension(
            'EXT_color_buffer_half_float'
          )
        }
        ;(this.vertexBuffer = xi(this.gl, this.debug)),
          (this.indexBuffer = bi(this.gl, this.debug)),
          (this.framebuffer = me(this.gl, this.debug)),
          (this.textureConfig = Qt(this.gl, this.textureHalfFloatExtension))
      }
      return (
        Object.defineProperty(t.prototype, 'debug', {
          get: function () {
            return i().getBool('DEBUG')
          },
          enumerable: !0,
          configurable: !0,
        }),
        (t.prototype.dispose = function () {
          var t = this
          if (!this.disposed) {
            null != this.program &&
              console.warn(
                'Disposing a GPGPUContext that still has a bound WebGLProgram. This is probably a resource leak, delete the program with GPGPUContext.deleteProgram before disposing.'
              ),
              null != this.outputTexture &&
                console.warn(
                  'Disposing a GPGPUContext that still has a bound output matrix texture.  This is probably a resource leak, delete the output matrix texture with GPGPUContext.deleteMatrixTexture before disposing.'
                )
            var e = this.gl
            Jt(e, this.debug, function () {
              return e.finish()
            }),
              Jt(e, this.debug, function () {
                return e.bindFramebuffer(e.FRAMEBUFFER, null)
              }),
              Jt(e, this.debug, function () {
                return e.deleteFramebuffer(t.framebuffer)
              }),
              Jt(e, this.debug, function () {
                return e.bindBuffer(e.ARRAY_BUFFER, null)
              }),
              Jt(e, this.debug, function () {
                return e.bindBuffer(e.ELEMENT_ARRAY_BUFFER, null)
              }),
              Jt(e, this.debug, function () {
                return e.deleteBuffer(t.indexBuffer)
              }),
              (this.disposed = !0)
          }
        }),
        (t.prototype.createFloat32MatrixTexture = function (t, e) {
          return (
            this.throwIfDisposed(),
            Ci(this.gl, this.debug, t, e, this.textureConfig)
          )
        }),
        (t.prototype.createFloat16MatrixTexture = function (t, e) {
          return (
            this.throwIfDisposed(),
            Ei(this.gl, this.debug, t, e, this.textureConfig)
          )
        }),
        (t.prototype.createUnsignedBytesMatrixTexture = function (t, e) {
          return (
            this.throwIfDisposed(),
            Ri(this.gl, this.debug, t, e, this.textureConfig)
          )
        }),
        (t.prototype.uploadPixelDataToTexture = function (t, e) {
          this.throwIfDisposed(), Di(this.gl, this.debug, t, e)
        }),
        (t.prototype.uploadDenseMatrixToTexture = function (t, e, n, r) {
          this.throwIfDisposed(),
            Ai(this.gl, this.debug, t, e, n, r, this.textureConfig)
        }),
        (t.prototype.createFloat16PackedMatrixTexture = function (t, e) {
          return (
            this.throwIfDisposed(),
            ki(this.gl, this.debug, t, e, this.textureConfig)
          )
        }),
        (t.prototype.createPackedMatrixTexture = function (t, e) {
          return (
            this.throwIfDisposed(),
            Ii(this.gl, this.debug, t, e, this.textureConfig)
          )
        }),
        (t.prototype.deleteMatrixTexture = function (t) {
          var e = this
          this.throwIfDisposed(),
            this.outputTexture === t &&
              (Ee(this.gl, this.debug, this.framebuffer),
              (this.outputTexture = null)),
            Jt(this.gl, this.debug, function () {
              return e.gl.deleteTexture(t)
            })
        }),
        (t.prototype.downloadByteEncodedFloatMatrixFromOutputTexture =
          function (t, e, n) {
            var r = this
            return this.downloadMatrixDriver(t, function () {
              return Fi(r.gl, r.debug, e, n, r.textureConfig)
            })
          }),
        (t.prototype.downloadPackedMatrixFromBuffer = function (
          t,
          e,
          n,
          r,
          o,
          a
        ) {
          return _i(this.gl, t, 0, 0, 0, o, a, this.textureConfig)
        }),
        (t.prototype.downloadFloat32MatrixFromBuffer = function (t, e) {
          return Ni(this.gl, t, e)
        }),
        (t.prototype.createBufferFromTexture = function (t, e, n) {
          this.bindTextureToFrameBuffer(t)
          var r = Ti(this.gl, this.debug, e, n, this.textureConfig)
          return this.unbindTextureToFrameBuffer(), r
        }),
        (t.prototype.createAndWaitForFence = function () {
          var t = this.createFence(this.gl)
          return this.pollFence(t)
        }),
        (t.prototype.createFence = function (t) {
          var e,
            n,
            r = this
          if (i().getBool('WEBGL_FENCE_API_ENABLED')) {
            var o = t,
              a = o.fenceSync(o.SYNC_GPU_COMMANDS_COMPLETE, 0)
            t.flush(),
              (n = function () {
                var t = o.clientWaitSync(a, 0, 0)
                return t === o.ALREADY_SIGNALED || t === o.CONDITION_SATISFIED
              }),
              (e = a)
          } else
            i().getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION') > 0
              ? ((e = this.beginQuery()),
                this.endQuery(),
                (n = function () {
                  return r.isQueryAvailable(
                    e,
                    i().getNumber(
                      'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION'
                    )
                  )
                }))
              : (n = function () {
                  return !0
                })
          return { query: e, isFencePassed: n }
        }),
        (t.prototype.downloadMatrixFromPackedTexture = function (t, e, n) {
          var r = this
          return this.downloadMatrixDriver(t, function () {
            return Oi(r.gl, r.debug, e, n)
          })
        }),
        (t.prototype.createProgram = function (t) {
          this.throwIfDisposed()
          var e = this.gl,
            n = ae(e, this.debug, t),
            r = yi(e, this.debug),
            o = ce(e, this.debug)
          return (
            Jt(e, this.debug, function () {
              return e.attachShader(o, r)
            }),
            Jt(e, this.debug, function () {
              return e.attachShader(o, n)
            }),
            le(e, this.debug, o),
            this.debug && he(e, this.debug, o),
            this.vertexAttrsAreBound ||
              (this.setProgram(o),
              (this.vertexAttrsAreBound = Si(
                e,
                this.debug,
                this.program,
                this.vertexBuffer
              ))),
            o
          )
        }),
        (t.prototype.deleteProgram = function (t) {
          var e = this
          this.throwIfDisposed(),
            t === this.program && (this.program = null),
            null != t &&
              Jt(this.gl, this.debug, function () {
                return e.gl.deleteProgram(t)
              })
        }),
        (t.prototype.setProgram = function (t) {
          var e = this
          this.throwIfDisposed(),
            (this.program = t),
            null != this.program &&
              this.debug &&
              he(this.gl, this.debug, this.program),
            Jt(this.gl, this.debug, function () {
              return e.gl.useProgram(t)
            })
        }),
        (t.prototype.getUniformLocation = function (t, e, n) {
          return (
            void 0 === n && (n = !0),
            this.throwIfDisposed(),
            n ? xe(this.gl, this.debug, t, e) : be(this.gl, t, e)
          )
        }),
        (t.prototype.getAttributeLocation = function (t, e) {
          var n = this
          return (
            this.throwIfDisposed(),
            Jt(this.gl, this.debug, function () {
              return n.gl.getAttribLocation(t, e)
            })
          )
        }),
        (t.prototype.getUniformLocationNoThrow = function (t, e) {
          return this.throwIfDisposed(), this.gl.getUniformLocation(t, e)
        }),
        (t.prototype.setInputMatrixTexture = function (t, e, n) {
          this.throwIfDisposed(),
            this.throwIfNoProgram(),
            we(this.gl, this.debug, this.program, t, e, n)
        }),
        (t.prototype.setOutputMatrixTexture = function (t, e, n) {
          this.setOutputMatrixTextureDriver(t, n, e)
        }),
        (t.prototype.setOutputPackedMatrixTexture = function (t, e, n) {
          this.throwIfDisposed()
          var r = $t(e, n),
            o = r[0],
            a = r[1]
          this.setOutputMatrixTextureDriver(t, o, a)
        }),
        (t.prototype.setOutputMatrixWriteRegion = function (t, e, n, r) {
          this.setOutputMatrixWriteRegionDriver(n, t, r, e)
        }),
        (t.prototype.setOutputPackedMatrixWriteRegion = function (t, e, n, r) {
          throw new Error('setOutputPackedMatrixWriteRegion not implemented.')
        }),
        (t.prototype.debugValidate = function () {
          null != this.program && he(this.gl, this.debug, this.program),
            Re(this.gl)
        }),
        (t.prototype.executeProgram = function () {
          this.throwIfDisposed(), this.throwIfNoProgram()
          var t = this.gl
          this.debug && this.debugValidate(),
            Jt(t, this.debug, function () {
              return t.drawElements(t.TRIANGLES, 6, t.UNSIGNED_SHORT, 0)
            })
        }),
        (t.prototype.blockUntilAllProgramsCompleted = function () {
          var t = this
          this.throwIfDisposed(),
            Jt(this.gl, this.debug, function () {
              return t.gl.finish()
            })
        }),
        (t.prototype.getQueryTimerExtension = function () {
          return (
            null == this.disjointQueryTimerExtension &&
              (this.disjointQueryTimerExtension = re(
                this.gl,
                this.debug,
                2 ===
                  i().getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION')
                  ? 'EXT_disjoint_timer_query_webgl2'
                  : 'EXT_disjoint_timer_query'
              )),
            this.disjointQueryTimerExtension
          )
        }),
        (t.prototype.getQueryTimerExtensionWebGL2 = function () {
          return this.getQueryTimerExtension()
        }),
        (t.prototype.getQueryTimerExtensionWebGL1 = function () {
          return this.getQueryTimerExtension()
        }),
        (t.prototype.beginQuery = function () {
          if (
            2 === i().getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION')
          ) {
            var t = this.gl,
              e = this.getQueryTimerExtensionWebGL2(),
              n = t.createQuery()
            return t.beginQuery(e.TIME_ELAPSED_EXT, n), n
          }
          var r = this.getQueryTimerExtensionWebGL1(),
            o = r.createQueryEXT()
          return r.beginQueryEXT(r.TIME_ELAPSED_EXT, o), o
        }),
        (t.prototype.endQuery = function () {
          if (
            2 !== i().getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION')
          ) {
            var t = this.getQueryTimerExtensionWebGL1()
            t.endQueryEXT(t.TIME_ELAPSED_EXT)
          } else {
            var e = this.gl,
              n = this.getQueryTimerExtensionWebGL2()
            e.endQuery(n.TIME_ELAPSED_EXT)
          }
        }),
        (t.prototype.waitForQueryAndGetTime = function (t) {
          return n(this, void 0, void 0, function () {
            var e = this
            return r(this, function (n) {
              switch (n.label) {
                case 0:
                  return [
                    4,
                    F(function () {
                      return (
                        e.disposed ||
                        e.isQueryAvailable(
                          t,
                          i().getNumber(
                            'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION'
                          )
                        )
                      )
                    }),
                  ]
                case 1:
                  return (
                    n.sent(),
                    [
                      2,
                      this.getQueryTime(
                        t,
                        i().getNumber(
                          'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION'
                        )
                      ),
                    ]
                  )
              }
            })
          })
        }),
        (t.prototype.getQueryTime = function (t, e) {
          if (0 === e) return null
          if (2 === e) {
            var n = this.gl
            return n.getQueryParameter(t, n.QUERY_RESULT) / 1e6
          }
          var r = this.getQueryTimerExtensionWebGL1()
          return r.getQueryObjectEXT(t, r.QUERY_RESULT_EXT) / 1e6
        }),
        (t.prototype.isQueryAvailable = function (t, e) {
          if (0 === e) return !0
          if (2 === e) {
            var n = this.gl,
              r = this.getQueryTimerExtensionWebGL2(),
              o = n.getQueryParameter(t, n.QUERY_RESULT_AVAILABLE)
            return (
              null == this.disjoint &&
                (this.disjoint = this.gl.getParameter(r.GPU_DISJOINT_EXT)),
              o && !this.disjoint
            )
          }
          o = (r = this.getQueryTimerExtensionWebGL1()).getQueryObjectEXT(
            t,
            r.QUERY_RESULT_AVAILABLE_EXT
          )
          return (
            null == this.disjoint &&
              (this.disjoint = this.gl.getParameter(r.GPU_DISJOINT_EXT)),
            o && !this.disjoint
          )
        }),
        (t.prototype.pollFence = function (t) {
          var e = this
          return new Promise(function (n) {
            e.addItemToPoll(
              function () {
                return t.isFencePassed()
              },
              function () {
                return n()
              }
            )
          })
        }),
        (t.prototype.pollItems = function () {
          for (
            var t = (function (t) {
                for (var e = 0; e < t.length; ++e) {
                  if (!t[e]()) break
                }
                return e - 1
              })(
                this.itemsToPoll.map(function (t) {
                  return t.isDoneFn
                })
              ),
              e = 0;
            e <= t;
            ++e
          ) {
            ;(0, this.itemsToPoll[e].resolveFn)()
          }
          this.itemsToPoll = this.itemsToPoll.slice(t + 1)
        }),
        (t.prototype.addItemToPoll = function (t, e) {
          var n = this
          this.itemsToPoll.push({ isDoneFn: t, resolveFn: e }),
            this.itemsToPoll.length > 1 ||
              F(function () {
                return n.pollItems(), 0 === n.itemsToPoll.length
              })
        }),
        (t.prototype.bindTextureToFrameBuffer = function (t) {
          this.throwIfDisposed(),
            Ce(this.gl, this.debug, t, this.framebuffer),
            this.debug && Re(this.gl)
        }),
        (t.prototype.unbindTextureToFrameBuffer = function () {
          null != this.outputTexture
            ? (Ce(this.gl, this.debug, this.outputTexture, this.framebuffer),
              this.debug && Re(this.gl))
            : Ee(this.gl, this.debug, this.framebuffer)
        }),
        (t.prototype.downloadMatrixDriver = function (t, e) {
          this.bindTextureToFrameBuffer(t)
          var n = e()
          return this.unbindTextureToFrameBuffer(), n
        }),
        (t.prototype.setOutputMatrixTextureDriver = function (t, e, n) {
          this.throwIfDisposed()
          var r = this.gl
          Ce(r, this.debug, t, this.framebuffer),
            this.debug && Re(r),
            (this.outputTexture = t),
            Jt(r, this.debug, function () {
              return r.viewport(0, 0, e, n)
            }),
            Jt(r, this.debug, function () {
              return r.scissor(0, 0, e, n)
            })
        }),
        (t.prototype.setOutputMatrixWriteRegionDriver = function (t, e, n, r) {
          var o = this
          this.throwIfDisposed(),
            Jt(this.gl, this.debug, function () {
              return o.gl.scissor(t, e, n, r)
            })
        }),
        (t.prototype.throwIfDisposed = function () {
          if (this.disposed)
            throw new Error('Attempted to use disposed GPGPUContext.')
        }),
        (t.prototype.throwIfNoProgram = function () {
          if (null == this.program)
            throw new Error('No GPU program is currently set.')
        }),
        t
      )
    })()
  function Pi(t, e) {
    if (t.length !== e.length)
      throw Error(
        'Binary was compiled with ' +
          t.length +
          ' inputs, but was executed with ' +
          e.length +
          ' inputs'
      )
    t.forEach(function (t, n) {
      var r = t.logicalShape,
        o = e[n],
        a = o.shape
      if (!S(r, a))
        throw Error(
          'Binary was compiled with different shapes than the current args. Shapes ' +
            r +
            ' and ' +
            a +
            ' must match'
        )
      if (!t.isUniform || !o.isUniform) {
        var i = t.texShape,
          s = o.isUniform ? null : o.texData.texShape
        if (!S(i, s))
          throw Error(
            'Binary was compiled with different texture shapes than the current args. Shape ' +
              i +
              ' and ' +
              s +
              ' must match'
          )
      }
    })
  }
  var Li = function (t, e, n) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = t)
      for (
        var r = n.filterWidth,
          o = n.inChannels,
          a = n.strideWidth,
          i = n.strideHeight,
          s = n.padInfo,
          u = n.outWidth,
          c = n.dilationWidth,
          l = n.dilationHeight,
          h = n.dataFormat,
          f = s.left,
          d = s.top,
          p = o * r,
          v = ua(),
          m = 'channelsLast' === h,
          g = m ? 0 : 1,
          y = m ? 1 : 2,
          x = '',
          b = 0;
        b <= 1;
        b++
      )
        for (var w = 0; w <= 1; w++)
          x +=
            '\n          blockIndex = rc.y + ' +
            w +
            ';\n          pos = rc.x + ' +
            b +
            ';\n\n          if(blockIndex < ' +
            t[1] +
            ' && pos < ' +
            t[0] +
            ') {\n            offsetY = int(blockIndex / (' +
            u +
            ')) * ' +
            i +
            ' - ' +
            d +
            ';\n            d0 = offsetY + ' +
            l +
            ' * (pos / ' +
            p +
            ');\n\n            if(d0 < ' +
            e[g] +
            ' && d0 >= 0) {\n\n              offsetX = int(mod(float(blockIndex), ' +
            u +
            '.) * ' +
            a +
            '. - ' +
            f +
            '.);\n              d1 = offsetX + ' +
            c +
            ' * (int(mod(float(pos), ' +
            p +
            '.) / ' +
            o +
            '.));\n\n              if(d1 < ' +
            e[y] +
            ' && d1 >= 0) {\n\n                ch = int(mod(float(pos), ' +
            o +
            '.));\n\n                if (' +
            m +
            ') {\n                  innerDims = vec2(d1, ch);\n                  result[' +
            (2 * b + w) +
            '] = getChannel(\n                    getA(d0, int(innerDims.x),\n                    int(innerDims.y)), innerDims);\n                } else {\n                  innerDims = vec2(d0, d1);\n                  result[' +
            (2 * b + w) +
            '] = getChannel(\n                    getA(ch, int(innerDims.x),\n                    int(innerDims.y)), innerDims);\n                }\n              }\n            }\n          }\n        '
      this.userCode =
        '\n      void main() {\n        ivec2 rc = getOutputCoords();\n\n        vec4 result = vec4(0);\n\n        int blockIndex, pos, offsetY, d0, offsetX, d1, ch;\n        vec2 innerDims;\n\n        ' +
        x +
        '\n\n        ' +
        v.output +
        ' = result;\n      }\n    '
    },
    Wi = function (t, e, n, r, o) {
      ;(this.variableNames = ['x']), (this.outputShape = [])
      var a,
        i = e,
        s = t[3] - 1
      this.outputShape = t
      var u = 'float(' + n + ') + float(' + r + ') * sum'
      ;(a =
        0.5 === o
          ? 'inversesqrt(' + u + ')'
          : 1 === o
          ? '1.0/(' + u + ')'
          : 'exp(log(' + u + ') * float(-' + o + '));'),
        (this.userCode =
          '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int r = coords[1];\n        int c = coords[2];\n        int d = coords[3];\n        float x = getX(b, r, c, d);\n        float sum = 0.0;\n        for (int j = -' +
          i +
          '; j <= ' +
          i +
          '; j++) {\n          int idx = d + j;\n          if (idx >= 0 && idx <=  ' +
          s +
          ') {\n            float z = getX(b, r, c, idx);\n            sum += z * z;\n          }\n        }\n        float val = x * ' +
          a +
          ';\n        setOutput(val);\n      }\n    ')
    },
    Ui = function (t, e, n, r, o) {
      ;(this.variableNames = ['inputImage', 'outputImage', 'dy']),
        (this.outputShape = []),
        (this.outputShape = t),
        (this.depth = t[3]),
        (this.depthRadius = e),
        (this.bias = n),
        (this.alpha = r),
        (this.beta = o),
        (this.userCode =
          '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int r = coords[1];\n        int c = coords[2];\n\n        float result = 0.0;\n        for (int d = 0; d < ' +
          this.depth +
          '; ++d) {\n          int depthBegin = int(max(0.0, float(d - ' +
          e +
          ')));\n          int depthEnd = int(min(float(' +
          this.depth +
          '),\n              float(d + ' +
          e +
          ' + 1)));\n\n          const int MIN_DEPTH_BEGIN = 0;\n          const int MAX_DEPTH_END = ' +
          this.depth +
          ';\n\n          float norm = 0.0;\n          for (int k = MIN_DEPTH_BEGIN; k < MAX_DEPTH_END; ++k) {\n            if (k < depthBegin){\n              continue;\n            }\n            else if (k >= depthBegin && k < depthEnd) {\n              norm += getInputImage(b, r, c, k) * getInputImage(b, r, c, k);\n            }\n            else {\n              break;\n            }\n          }\n\n          norm = float(' +
          r +
          ') * norm + float(' +
          n +
          ');\n\n          for(int k = MIN_DEPTH_BEGIN; k < MAX_DEPTH_END; ++k){\n            if (k < depthBegin){\n              continue;\n            }\n            else if (k >= depthBegin && k < depthEnd){\n              float dyi = -2.0 * float(' +
          r +
          ')\n                * float(' +
          o +
          ')\n                * getInputImage(b ,r ,c, k) * getOutputImage(b, r, c, d)\n                / norm;\n              if (k == d) {\n                dyi += pow(norm, -1.0 * ' +
          o +
          ');\n              }\n              if (k == coords[3]) {\n                dyi *= getDy(b, r, c, d);\n                result += dyi;\n              }\n            }\n            else {\n              break;\n            }\n          }\n      }\n      setOutput(result);\n      }\n    ')
    },
    Vi = function (t, e, n, r, o) {
      ;(this.variableNames = ['x']),
        (this.outputShape = []),
        (this.packedInputs = !0),
        (this.packedOutput = !0)
      var a,
        i = e,
        s = t[3] - 1
      this.outputShape = t
      var u = 'float(' + n + ') + float(' + r + ') * sum'
      ;(a =
        0.5 === o
          ? 'inversesqrt(' + u + ')'
          : 1 === o
          ? '1.0/(' + u + ')'
          : 'exp(log(' + u + ') * float(-' + o + '));'),
        (this.userCode =
          '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords.x;\n        int r = coords.y;\n        int c = coords.z;\n        int d = coords.w;\n\n        bool hasNextCol = d < ' +
          this.outputShape[3] +
          ';\n        bool hasNextRow = c < ' +
          this.outputShape[2] +
          ';\n\n        vec4 sum = vec4(0.);\n        vec4 xFragAtOutputCoords = getX(b, r, c, d);\n\n        vec4 xAtOutputCoords = vec4(\n          getChannel(xFragAtOutputCoords, vec2(c, d)),\n          hasNextCol ?\n            getChannel(xFragAtOutputCoords, vec2(c, d + 1)) : 0.0,\n          hasNextRow ?\n            getChannel(xFragAtOutputCoords , vec2(c + 1, d)) : 0.0,\n          (hasNextRow && hasNextCol) ?\n            getChannel(xFragAtOutputCoords, vec2(c + 1, d + 1)) : 0.0\n        );\n\n        int firstChannel = d - ' +
          i +
          ';\n        vec2 cache = vec2(0.);\n        if(firstChannel >= 0){\n          vec4 firstChannelFrag = getX(b, r, c, firstChannel);\n          cache.x = getChannel(firstChannelFrag, vec2(c, firstChannel));\n            if(hasNextRow){\n              cache.y = getChannel(firstChannelFrag, vec2(c + 1, firstChannel));\n            }\n        }\n\n        ivec2 depth = ivec2(d, d + 1);\n        for (int j = - ' +
          i +
          '; j <= ' +
          i +
          '; j++) {\n          ivec2 idx = depth + j;\n          bvec2 aboveLowerBound = greaterThanEqual(idx, ivec2(0));\n          bvec2 belowUpperBound = lessThanEqual(idx, ivec2(' +
          s +
          '));\n\n          bool depthInRange = aboveLowerBound.x && belowUpperBound.x;\n          bool depthPlusOneInRange = aboveLowerBound.y && belowUpperBound.y;\n\n          if(depthInRange || depthPlusOneInRange){\n            vec4 z = vec4(0.);\n            vec4 xFragAtCurrentDepth;\n            z.xz = cache.xy;\n            if(depthPlusOneInRange && hasNextCol){\n              xFragAtCurrentDepth = idx.y != d ?\n                getX(b, r, c, idx.y) : xFragAtOutputCoords;\n              z.y = getChannel(xFragAtCurrentDepth, vec2(c, idx.y));\n              if(hasNextRow){\n                z.w = getChannel(xFragAtCurrentDepth, vec2(c + 1, idx.y));\n              }\n            }\n            cache.xy = z.yw;\n            sum += z * z;\n          }\n        }\n        vec4 result = xAtOutputCoords * ' +
          a +
          ';\n        setOutput(result);\n      }\n    ')
    },
    zi = function (t) {
      ;(this.variableNames = ['dy', 'maxPos']), (this.outputShape = t.inShape)
      var e = t.strideHeight,
        n = t.strideWidth,
        r = t.dilationHeight,
        o = t.effectiveFilterHeight,
        a = t.effectiveFilterWidth,
        i = o - 1 - t.padInfo.top,
        s = a - 1 - t.padInfo.left,
        u = o * a - 1
      this.userCode =
        '\n      const ivec2 pads = ivec2(' +
        i +
        ', ' +
        s +
        ');\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n\n        ivec2 dyRCCorner = coords.yz - pads;\n        int dyRCorner = dyRCCorner.x;\n        int dyCCorner = dyRCCorner.y;\n\n        // Convolve dy(?, ?, d) with pos mask(:, :, d) to get dx(xR, xC, d).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n        for (int wR = 0; wR < ' +
        o +
        ';\n          wR += ' +
        r +
        ') {\n          float dyR = float(dyRCorner + wR) / ' +
        e +
        '.0;\n\n          if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 || fract(dyR) > 0.0) {\n            continue;\n          }\n          int idyR = int(dyR);\n\n          for (int wC = 0; wC < ' +
        a +
        '; wC++) {\n            float dyC = float(dyCCorner + wC) / ' +
        n +
        '.0;\n\n            if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                fract(dyC) > 0.0) {\n              continue;\n            }\n            int idyC = int(dyC);\n\n            float dyValue = getDy(b, idyR, idyC, d);\n            int maxPosValue = ' +
        u +
        ' - int(getMaxPos(b, idyR, idyC, d));\n\n            // Get the current value, check it against the value from the\n            // position matrix.\n            int curPosValue = wR * ' +
        a +
        ' + wC;\n            float mask = float(maxPosValue == curPosValue ? 1.0 : 0.0);\n\n            dotProd += dyValue * mask;\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Gi = function (t) {
      ;(this.variableNames = ['dy', 'maxPos']), (this.outputShape = t.inShape)
      var e = t.strideDepth,
        n = t.strideHeight,
        r = t.strideWidth,
        o = t.dilationDepth,
        a = t.dilationHeight,
        i = t.dilationWidth,
        s = t.effectiveFilterDepth,
        u = t.effectiveFilterHeight,
        c = t.effectiveFilterWidth,
        l = s - 1 - t.padInfo.front,
        h = u - 1 - t.padInfo.top,
        f = c - 1 - t.padInfo.left,
        d = s * u * c - 1
      this.userCode =
        '\n      const ivec3 pads = ivec3(' +
        l +
        ', ' +
        h +
        ', ' +
        f +
        ');\n\n      void main() {\n        ivec5 coords = getOutputCoords();\n        int batch = coords.x;\n        int ch = coords.u;\n\n        ivec3 dyCorner = ivec3(coords.y, coords.z, coords.w) - pads;\n        int dyDCorner = dyCorner.x;\n        int dyRCorner = dyCorner.y;\n        int dyCCorner = dyCorner.z;\n\n        // Convolve dy(?, ?, ?, ch) with pos mask(:, :, :, d) to get\n        // dx(xD, xR, xC, ch).\n        // ? = to be determined. : = across all values in that axis.\n        float dotProd = 0.0;\n\n        for (int wD = 0; wD < ' +
        s +
        ';\n           wD += ' +
        o +
        ') {\n          float dyD = float(dyDCorner + wD) / ' +
        e +
        '.0;\n\n          if (dyD < 0.0 || dyD >= ' +
        t.outDepth +
        '.0 || fract(dyD) > 0.0) {\n            continue;\n          }\n          int idyD = int(dyD);\n\n          for (int wR = 0; wR < ' +
        u +
        ';\n              wR += ' +
        a +
        ') {\n            float dyR = float(dyRCorner + wR) / ' +
        n +
        '.0;\n\n            if (dyR < 0.0 || dyR >= ' +
        t.outHeight +
        '.0 ||\n                fract(dyR) > 0.0) {\n              continue;\n            }\n            int idyR = int(dyR);\n\n            for (int wC = 0; wC < ' +
        c +
        ';\n                wC += ' +
        i +
        ') {\n              float dyC = float(dyCCorner + wC) / ' +
        r +
        '.0;\n\n              if (dyC < 0.0 || dyC >= ' +
        t.outWidth +
        '.0 ||\n                  fract(dyC) > 0.0) {\n                continue;\n              }\n              int idyC = int(dyC);\n\n              float dyValue = getDy(batch, idyD, idyR, idyC, ch);\n              int maxPosValue = ' +
        d +
        ' -\n                  int(getMaxPos(batch, idyD, idyR, idyC, ch));\n\n              // Get the current value, check it against the value from the\n              // position matrix.\n              int curPosValue =\n                  wD * ' +
        u +
        ' * ' +
        c +
        ' +\n                  wR * ' +
        c +
        ' + wC;\n              float mask = float(maxPosValue == curPosValue ? 1.0 : 0.0);\n\n              dotProd += dyValue * mask;\n            }\n          }\n        }\n        setOutput(dotProd);\n      }\n    '
    },
    Hi = function (t, e, n, r, o, a, i) {
      void 0 === n && (n = !1),
        void 0 === r && (r = !1),
        void 0 === o && (o = !1),
        void 0 === a && (a = null),
        void 0 === i && (i = !1),
        (this.variableNames = ['matrixA', 'matrixB']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = e)
      var s = n ? t[1] : t[2],
        u = Math.ceil(s / 2),
        c = n ? 'i * 2, rc.y' : 'rc.y, i * 2',
        l = r ? 'rc.z, i * 2' : 'i * 2, rc.z',
        h = n ? ['a.xxyy', 'a.zzww'] : ['a.xxzz', 'a.yyww'],
        f = r ? ['b.xzxz', 'b.ywyw'] : ['b.xyxy', 'b.zwzw'],
        d = '',
        p = ''
      a &&
        ((d = i
          ? 'vec4 activation(vec4 a) {\n          vec4 b = getPreluActivationWeightsAtOutCoords();\n          ' +
            a +
            '\n        }'
          : 'vec4 activation(vec4 x) {\n          ' + a + '\n        }'),
        (p = 'result = activation(result);'))
      var v = o ? 'result += getBiasAtOutCoords();' : ''
      o && this.variableNames.push('bias'),
        i && this.variableNames.push('preluActivationWeights'),
        (this.userCode =
          '\n      ' +
          d +
          '\n\n      const float sharedDimension = ' +
          u +
          '.0;\n\n      vec4 dot2x2ARowBCol(ivec3 rc) {\n        vec4 result = vec4(0);\n        for (int i = 0; i < ' +
          u +
          '; i++) {\n          vec4 a = getMatrixA(rc.x, ' +
          c +
          ');\n          vec4 b = getMatrixB(rc.x, ' +
          l +
          ');\n\n          // These swizzled products need to be separately added.\n          // See: https://github.com/tensorflow/tfjs/issues/1735\n          result += (' +
          h[0] +
          ' * ' +
          f[0] +
          ');\n          result += (' +
          h[1] +
          ' * ' +
          f[1] +
          ');\n        }\n        return result;\n      }\n\n      void main() {\n        ivec3 rc = getOutputCoords();\n        vec4 result = dot2x2ARowBCol(rc);\n\n        ' +
          v +
          '\n\n        ' +
          p +
          '\n\n        setOutput(result);\n      }\n    ')
    },
    qi = (function () {
      function t(t, e, n) {
        ;(this.variableNames = ['probs']),
          (this.outputShape = [t, n]),
          (this.userCode =
            '\n      uniform float seed;\n\n      void main() {\n        ivec2 coords = getOutputCoords();\n        int batch = coords[0];\n\n        float r = random(seed);\n        float cdf = 0.0;\n\n        for (int i = 0; i < ' +
            (e - 1) +
            '; i++) {\n          cdf += getProbs(batch, i);\n\n          if (r < cdf) {\n            setOutput(float(i));\n            return;\n          }\n        }\n\n        // If no other event happened, last event happened.\n        setOutput(float(' +
            (e - 1) +
            '));\n      }\n    ')
      }
      return (
        (t.prototype.getCustomSetupFunc = function (t) {
          var e = this
          return function (n, r) {
            null == e.seedLoc && (e.seedLoc = n.getUniformLocation(r, 'seed')),
              n.gl.uniform1f(e.seedLoc, t)
          }
        }),
        t
      )
    })(),
    Ki = function (t, e, n, r) {
      ;(this.variableNames = ['indices']),
        (this.outputShape = [t, e]),
        (this.userCode =
          '\n      void main() {\n        ivec2 coords = getOutputCoords();\n        int index = round(getIndices(coords.x));\n        setOutput(mix(float(' +
          r +
          '), float(' +
          n +
          '),\n                      float(index == coords.y)));\n      }\n    ')
    },
    ji = function (t) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !1),
        (this.packedOutput = !0),
        (this.outputShape = t)
      var e = t.length
      if (0 === e)
        this.userCode =
          '\n        void main() {\n          setOutput(vec4(getA(), 0., 0., 0.));\n        }\n      '
      else {
        var n = sa('rc', e),
          r = wa(e),
          o = (function (t, e, n) {
            if (1 === t) return 'rc > ' + e[0]
            for (var r = '', o = t - 2; o < t; o++)
              (r += n[o] + ' >= ' + e[o]), o < t - 1 && (r += '||')
            return r
          })(e, t, n),
          a = (function (t, e, n, r) {
            if (1 === t) return ''
            var o = r.slice(-2)
            return (
              '\n    int r = ' +
              o[0] +
              ';\n    int c = ' +
              o[1] +
              ';\n    int rp1 = r + 1;\n    int cp1 = c + 1;\n\n    bool cEdge = cp1 >= ' +
              e +
              ';\n    bool rEdge = rp1 >= ' +
              n +
              ';\n  '
            )
          })(e, t[t.length - 1], t[t.length - 2], n),
          i = (function (t, e) {
            var n = t.length,
              r = (function (t, e) {
                for (var n = [], r = 0; r <= 1; r++)
                  for (var o = 0; o <= 1; o++) {
                    for (
                      var a =
                          (0 === r ? 'r' : 'rp1') +
                          ', ' +
                          (0 === o ? 'c' : 'cp1'),
                        i = 2;
                      i < t;
                      i++
                    )
                      a = e[e.length - 1 - i] + ',' + a
                    n.push(a)
                  }
                return n
              })(n, e)
            return 1 === n
              ? 'getA(rc),\n            rc + 1 >= ' +
                  t[0] +
                  ' ? 0. : getA(rc + 1),\n            0, 0'
              : 'getA(' +
                  r[0] +
                  '),\n          cEdge ? 0. : getA(' +
                  r[1] +
                  '),\n          rEdge ? 0. : getA(' +
                  r[2] +
                  '),\n          rEdge || cEdge ? 0. : getA(' +
                  r[3] +
                  ')'
          })(t, n)
        this.userCode =
          '\n        void main() {\n          ' +
          r +
          ' rc = getOutputCoords();\n\n          if(' +
          o +
          ') {\n            setOutput(vec4(0));\n          } else {\n            ' +
          a +
          '\n\n            setOutput(vec4(' +
          i +
          '));\n          }\n        }\n      '
      }
    }
  var Xi = function (t, e, n) {
      ;(this.variableNames = ['x']),
        (this.outputShape = e.map(function (e, n) {
          return e[0] + t[n] + e[1]
        }))
      var r = t.length,
        o = wa(r),
        a = e
          .map(function (t) {
            return t[0]
          })
          .join(','),
        i = e
          .map(function (e, n) {
            return e[0] + t[n]
          })
          .join(','),
        s = ['coords[0]', 'coords[1]', 'coords[2]', 'coords[3]'].slice(0, r)
      this.userCode =
        1 !== r
          ? '\n      ' +
            o +
            ' start = ' +
            o +
            '(' +
            a +
            ');\n      ' +
            o +
            ' end = ' +
            o +
            '(' +
            i +
            ');\n\n      void main() {\n        ' +
            o +
            ' outC = getOutputCoords();\n        if (any(lessThan(outC, start)) || any(greaterThanEqual(outC, end))) {\n          setOutput(float(' +
            n +
            '));\n        } else {\n          ' +
            o +
            ' coords = outC - start;\n          setOutput(getX(' +
            s +
            '));\n        }\n      }\n    '
          : '\n        int start = ' +
            a +
            ';\n        int end = ' +
            i +
            ';\n\n        void main() {\n          int outC = getOutputCoords();\n          if (outC < start || outC >= end) {\n            setOutput(float(' +
            n +
            '));\n          } else {\n            setOutput(getX(outC - start));\n          }\n        }\n      '
    },
    Yi = function (t, e, n) {
      ;(this.variableNames = ['x']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = e.map(function (e, n) {
          return e[0] + t[n] + e[1]
        }))
      for (
        var r = t.length,
          o = wa(r),
          a = e
            .map(function (t) {
              return t[0]
            })
            .join(','),
          i = e
            .map(function (e, n) {
              return e[0] + t[n]
            })
            .join(','),
          s = sa('rc', r),
          u = sa('source', r),
          c = s[r - 1] + ' < ' + this.outputShape[r - 1],
          l = 1 === r ? 'source' : 'vec2(' + u.slice(-2).join() + ')',
          h = [
            o + ' rc = outputLoc;',
            s[r - 1] + ' += 1;\n       if(' + c + ') {\n      ',
            1 === r
              ? ''
              : '}\n       rc = outputLoc;\n       ' +
                s[r - 2] +
                ' += 1;\n       if(' +
                s[r - 2] +
                ' < ' +
                this.outputShape[r - 2] +
                ') {',
            1 === r ? '' : '  ' + s[r - 1] + ' += 1;\n         if(' + c + ') {',
          ],
          f =
            1 === r
              ? 'rc < start || rc >= end'
              : 'any(lessThan(rc, start)) || any(greaterThanEqual(rc, end))',
          d = '',
          p = 0,
          v = 1 === r ? 2 : 4;
        p < v;
        p++
      )
        d +=
          '\n        ' +
          h[p] +
          '\n        if (' +
          f +
          ') {\n          result[' +
          p +
          '] = float(' +
          n +
          ');\n        } else {\n          ' +
          o +
          ' source = rc - start;\n          result[' +
          p +
          '] = getChannel(getX(' +
          u.join() +
          '), ' +
          l +
          ');\n        }\n      '
      ;(d += 1 === r ? '} ' : '}}'),
        (this.userCode =
          '\n      const ' +
          o +
          ' start = ' +
          o +
          '(' +
          a +
          ');\n      const ' +
          o +
          ' end = ' +
          o +
          '(' +
          i +
          ');\n\n      void main() {\n        ' +
          o +
          ' outputLoc = getOutputCoords();\n        vec4 result = vec4(0.);\n        ' +
          d +
          '\n        setOutput(result);\n      }\n    ')
    },
    $i = function (t, e, n) {
      if (((this.variableNames = ['x']), 'avg' === e && n))
        throw new Error('Cannot compute positions for average pool.')
      var r = t.filterWidth,
        o = t.strideHeight,
        a = t.strideWidth,
        i = t.dilationHeight,
        s = t.dilationWidth,
        u = t.effectiveFilterHeight,
        c = t.effectiveFilterWidth,
        l = t.padInfo.top,
        h = t.padInfo.left
      this.outputShape = t.outShape
      var f = 'avg' === e,
        d = '0.0'
      if ((f || (d = '-1.0 / 1e-20'), n))
        this.userCode =
          '\n        const ivec2 strides = ivec2(' +
          o +
          ', ' +
          a +
          ');\n        const ivec2 pads = ivec2(' +
          l +
          ', ' +
          h +
          ');\n\n        void main() {\n          ivec4 coords = getOutputCoords();\n          int batch = coords[0];\n          int d = coords[3];\n\n          ivec2 xRCCorner = coords.yz * strides - pads;\n          int xRCorner = xRCCorner.x;\n          int xCCorner = xRCCorner.y;\n\n          // max/min x(?, ?, d) to get y(yR, yC, d).\n          // ? = to be determined\n          float minMaxValue = 0.0;\n          float minMaxValueFound = 0.0;\n          int minMaxPosition = 0;\n          float avgValue = 0.0;\n\n          for (int wR = 0; wR < ' +
          u +
          ';\n              wR += ' +
          i +
          ') {\n            int xR = xRCorner + wR;\n\n            if (xR < 0 || xR >= ' +
          t.inHeight +
          ') {\n              continue;\n            }\n\n            for (int wC = 0; wC < ' +
          c +
          ';\n                wC += ' +
          s +
          ') {\n              int xC = xCCorner + wC;\n\n              if (xC < 0 || xC >= ' +
          t.inWidth +
          ') {\n                continue;\n              }\n\n              float value = getX(batch, xR, xC, d);\n\n              // If a min / max value has already been found, use it. If not,\n              // use the current value.\n              float currMinMaxValue = mix(\n                  value, minMaxValue, minMaxValueFound);\n              if (value >= currMinMaxValue) {\n                minMaxValue = value;\n                minMaxValueFound = 1.0;\n                minMaxPosition = wR * ' +
          c +
          ' + wC;\n              }\n            }\n          }\n          setOutput(float(minMaxPosition));\n        }\n      '
      else {
        var p =
          e +
          '(' +
          e +
          '(' +
          e +
          '(minMaxValue[0], minMaxValue[1]), minMaxValue[2]), minMaxValue[3])'
        'avg' === e && (p = 'avgValue / count')
        var v = 4 * Math.floor(r / 4),
          m = r % 4,
          g =
            '\n      if (' +
            f +
            ') {\n        avgValue += dot(values, ones);\n      } else {\n        minMaxValue = max(values, minMaxValue);\n      }\n    '
        this.userCode =
          '\n      const ivec2 strides = ivec2(' +
          o +
          ', ' +
          a +
          ');\n      const ivec2 pads = ivec2(' +
          l +
          ', ' +
          h +
          ');\n      const float initializationValue = ' +
          d +
          ';\n      const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);\n\n      float count = 0.0;\n\n      float getValue(int batch, int xR, int xC, int d) {\n        if (xC < 0 || xC >= ' +
          t.inWidth +
          ') {\n          return initializationValue;\n        }\n        count += 1.0;\n        return getX(batch, xR, xC, d);\n      }\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int batch = coords[0];\n        int d = coords[3];\n\n        ivec2 xRCCorner = coords.yz * strides - pads;\n        int xRCorner = xRCCorner.x;\n        int xCCorner = xRCCorner.y;\n\n        // max/min x(?, ?, d) to get y(yR, yC, d).\n        // ? = to be determined\n        vec4 minMaxValue = vec4(' +
          d +
          ');\n        float avgValue = 0.0;\n        count = 0.0;\n\n        for (int wR = 0; wR < ' +
          u +
          ';\n            wR += ' +
          i +
          ') {\n          int xR = xRCorner + wR;\n\n          if (xR < 0 || xR >= ' +
          t.inHeight +
          ') {\n            continue;\n          }\n\n          for (int wC = 0; wC < ' +
          v +
          '; wC += 4) {\n            int xC = xCCorner + wC * ' +
          s +
          ';\n\n            vec4 values = vec4(\n              getValue(batch, xR, xC, d),\n              getValue(batch, xR, xC + ' +
          s +
          ', d),\n              getValue(batch, xR, xC + 2 * ' +
          s +
          ', d),\n              getValue(batch, xR, xC + 3 * ' +
          s +
          ', d)\n            );\n\n            ' +
          g +
          '\n          }\n\n          int xC = xCCorner + ' +
          v +
          ';\n          if (' +
          (1 === m) +
          ') {\n            vec4 values = vec4(\n              getValue(batch, xR, xC, d),\n              initializationValue,\n              initializationValue,\n              initializationValue\n            );\n\n            ' +
          g +
          '\n          } else if (' +
          (2 === m) +
          ') {\n            vec4 values = vec4(\n              getValue(batch, xR, xC, d),\n              getValue(batch, xR, xC + ' +
          s +
          ', d),\n              initializationValue,\n              initializationValue\n            );\n\n            ' +
          g +
          '\n          } else if (' +
          (3 === m) +
          ') {\n            vec4 values = vec4(\n              getValue(batch, xR, xC, d),\n              getValue(batch, xR, xC + ' +
          s +
          ', d),\n              getValue(batch, xR, xC + 2 * ' +
          s +
          ', d),\n              initializationValue\n            );\n\n            ' +
          g +
          '\n          }\n        }\n        setOutput(' +
          p +
          ');\n      }\n    '
      }
    },
    Qi = function (t, e, n) {
      if (((this.variableNames = ['x']), 'avg' === e && n))
        throw new Error('Cannot compute positions for average pool.')
      var r = t.filterWidth,
        o = t.strideDepth,
        a = t.strideHeight,
        i = t.strideWidth,
        s = t.dilationDepth,
        u = t.dilationHeight,
        c = t.dilationWidth,
        l = t.effectiveFilterDepth,
        h = t.effectiveFilterHeight,
        f = t.effectiveFilterWidth,
        d = t.padInfo.front,
        p = t.padInfo.top,
        v = t.padInfo.left
      this.outputShape = t.outShape
      var m = 'avg' === e,
        g = '0.0'
      if ((m || (g = '-1.0 / 1e-20'), n))
        this.userCode =
          '\n        const ivec3 strides =\n            ivec3(' +
          o +
          ', ' +
          a +
          ', ' +
          i +
          ');\n        const ivec3 pads = ivec3(' +
          d +
          ', ' +
          p +
          ', ' +
          v +
          ');\n\n        void main() {\n          ivec5 coords = getOutputCoords();\n          int batch = coords.x;\n          int ch = coords.u;\n\n          ivec3 xCorner = ivec3(coords.y, coords.z, coords.w) * strides - pads;\n          int xDCorner = xCorner.x;\n          int xRCorner = xCorner.y;\n          int xCCorner = xCorner.z;\n\n          // max/min x(?, ?, ?, ch) to get y(yD, yR, yC, ch).\n          // ? = to be determined\n          float minMaxValue = 0.0;\n          float minMaxValueFound = 0.0;\n          int minMaxPosition = 0;\n\n          for (int wD = 0; wD < ' +
          l +
          ';\n              wD += ' +
          s +
          ') {\n            int xD = xDCorner + wD;\n\n            if (xD < 0 || xD >= ' +
          t.inDepth +
          ') {\n              continue;\n            }\n\n            for (int wR = 0; wR < ' +
          h +
          ';\n                wR += ' +
          u +
          ') {\n              int xR = xRCorner + wR;\n\n              if (xR < 0 || xR >= ' +
          t.inHeight +
          ') {\n                continue;\n              }\n\n              for (int wC = 0; wC < ' +
          f +
          ';\n                  wC += ' +
          c +
          ') {\n                int xC = xCCorner + wC;\n\n                if (xC < 0 || xC >= ' +
          t.inWidth +
          ') {\n                  continue;\n                }\n\n                float value = getX(batch, xD, xR, xC, ch);\n\n                // If a min / max value has already been found, use it. If not,\n                // use the current value.\n                float currMinMaxValue = mix(\n                    value, minMaxValue, minMaxValueFound);\n                if (value >= currMinMaxValue) {\n                  minMaxValue = value;\n                  minMaxValueFound = 1.0;\n                  minMaxPosition =\n                      wD * ' +
          h +
          ' * ' +
          f +
          ' +\n                      wR * ' +
          f +
          ' + wC;;\n                }\n              }\n            }\n          }\n          setOutput(float(minMaxPosition));\n        }\n      '
      else {
        var y =
          e +
          '(' +
          e +
          '(' +
          e +
          '(minMaxValue[0], minMaxValue[1]), minMaxValue[2]), minMaxValue[3])'
        'avg' === e && (y = 'avgValue / count')
        var x = 4 * Math.floor(r / 4),
          b = r % 4,
          w =
            '\n      if (' +
            m +
            ') {\n        avgValue += dot(values, ones);\n      } else {\n        minMaxValue = max(values, minMaxValue);\n      }\n    '
        this.userCode =
          '\n      const ivec3 strides =\n        ivec3(' +
          o +
          ', ' +
          a +
          ', ' +
          i +
          ');\n      const ivec3 pads = ivec3(' +
          d +
          ', ' +
          p +
          ', ' +
          v +
          ');\n      const float initializationValue = ' +
          g +
          ';\n      const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);\n\n      float count = 0.0;\n\n      float getValue(int batch, int xD, int xR, int xC, int ch) {\n        if (xC < 0 || xC >= ' +
          t.inWidth +
          ') {\n          return initializationValue;\n        }\n        count += 1.0;\n        return getX(batch, xD, xR, xC, ch);\n      }\n\n      void main() {\n        ivec5 coords = getOutputCoords();\n        int batch = coords.x;\n        int ch = coords.u;\n\n        ivec3 xCorner = ivec3(coords.y, coords.z, coords.w) * strides - pads;\n        int xDCorner = xCorner.x;\n        int xRCorner = xCorner.y;\n        int xCCorner = xCorner.z;\n\n        // max/min x(?, ?, ?, d) to get y(yD, yR, yC, ch).\n        // ? = to be determined\n        vec4 minMaxValue = vec4(' +
          g +
          ');\n        float avgValue = 0.0;\n        count = 0.0;\n\n        for (int wD = 0; wD < ' +
          l +
          ';\n            wD += ' +
          s +
          ') {\n          int xD = xDCorner + wD;\n\n          if (xD < 0 || xD >= ' +
          t.inDepth +
          ') {\n            continue;\n          }\n\n          for (int wR = 0; wR < ' +
          h +
          ';\n            wR += ' +
          u +
          ') {\n            int xR = xRCorner + wR;\n\n            if (xR < 0 || xR >= ' +
          t.inHeight +
          ') {\n              continue;\n            }\n\n            for (int wC = 0; wC < ' +
          x +
          '; wC += 4) {\n              int xC = xCCorner + wC * ' +
          c +
          ';\n\n              vec4 values = vec4(\n                getValue(batch, xD, xR, xC, ch),\n                getValue(batch, xD, xR, xC + ' +
          c +
          ', ch),\n                getValue(batch, xD, xR, xC + 2 * ' +
          c +
          ', ch),\n                getValue(batch, xD, xR, xC + 3 * ' +
          c +
          ', ch)\n              );\n\n              ' +
          w +
          '\n            }\n\n            int xC = xCCorner + ' +
          x +
          ';\n            if (' +
          (1 === b) +
          ') {\n              vec4 values = vec4(\n                getValue(batch, xD, xR, xC, ch),\n                initializationValue,\n                initializationValue,\n                initializationValue\n              );\n\n              ' +
          w +
          '\n            } else if (' +
          (2 === b) +
          ') {\n              vec4 values = vec4(\n                getValue(batch, xD, xR, xC, ch),\n                getValue(batch, xD, xR, xC + ' +
          c +
          ', ch),\n                initializationValue,\n                initializationValue\n              );\n\n              ' +
          w +
          '\n            } else if (' +
          (3 === b) +
          ') {\n              vec4 values = vec4(\n                getValue(batch, xD, xR, xC, ch),\n                getValue(batch, xD, xR, xC + ' +
          c +
          ', ch),\n                getValue(batch, xD, xR, xC + 2 * ' +
          c +
          ', ch),\n                initializationValue\n              );\n\n              ' +
          w +
          '\n            }\n          }\n          setOutput(' +
          y +
          ');\n        }\n      }\n    '
      }
    },
    Ji = function (t, e) {
      this.variableNames = ['x']
      var n = t.windowSize,
        r = t.batchSize,
        o = t.inSize,
        a = Math.ceil(o / n)
      this.outputShape = [r, a]
      var i = '0.0',
        s = ''
      'prod' === e
        ? (i = '1.0')
        : 'min' === e
        ? ((i = '1.0 / 1e-20'), (s = 'min'))
        : 'max' === e && ((i = '-1.0 / 1e-20'), (s = 'max'))
      var u =
        e +
        '(' +
        e +
        '(' +
        e +
        '(minMaxValue[0], minMaxValue[1]), minMaxValue[2]), minMaxValue[3])'
      'sum' === e
        ? (u = 'sumValue')
        : 'prod' === e
        ? (u = 'prodValue')
        : 'all' === e
        ? (u = 'allValue')
        : 'any' === e && (u = 'anyValue')
      var c = 4 * Math.floor(n / 4),
        l = n % 4,
        h =
          '\n      if (' +
          ('sum' === e) +
          ') {\n        sumValue += dot(values, ones);\n      } else if (' +
          ('prod' === e) +
          ') {\n        vec2 tmp = vec2(values[0], values[1]) * vec2(values[2], values[3]);\n        prodValue *= tmp[0] * tmp[1];\n      } else {\n        minMaxValue = ' +
          s +
          '(values, minMaxValue);\n      }\n    ',
        f = 'vec4'
      'all' === e
        ? ((i = '1.0'),
          (h =
            '\n        bool reducedAllValue = all(values);\n        float floatedReducedAllValue = float(reducedAllValue);\n        allValue = float(allValue >= 1.0 && floatedReducedAllValue >= 1.0);\n      '),
          (f = 'bvec4'))
        : 'any' === e &&
          ((i = '0.0'),
          (h =
            '\n        bool reducedAnyValue = any(values);\n        float floatedReducedAnyValue = float(reducedAnyValue);\n        anyValue = float(anyValue >= 1.0 || floatedReducedAnyValue >= 1.0);\n      '),
          (f = 'bvec4'))
      var d = ''
      o % n > 0 &&
        (d =
          '\n        if (inIdx < 0 || inIdx >= ' +
          o +
          ') {\n          return initializationValue;\n        }\n      '),
        (this.userCode =
          '\n      const float initializationValue = ' +
          i +
          ';\n      const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);\n\n      float getValue(int batch, int inIdx) {\n        ' +
          d +
          '\n        return getX(batch, inIdx);\n      }\n\n      void main() {\n        ivec2 coords = getOutputCoords();\n        int batch = coords[0];\n        int outIdx = coords[1];\n        int inOffset = outIdx * ' +
          n +
          ';\n\n        vec4 minMaxValue = vec4(' +
          i +
          ');\n        float prodValue = 1.0;\n        float sumValue = 0.0;\n        float allValue = 1.0;\n        float anyValue = 0.0;\n\n        for (int i = 0; i < ' +
          c +
          '; i += 4) {\n          int inIdx = inOffset + i;\n          ' +
          f +
          ' values = ' +
          f +
          '(\n            getValue(batch, inIdx),\n            getValue(batch, inIdx + 1),\n            getValue(batch, inIdx + 2),\n            getValue(batch, inIdx + 3)\n          );\n\n          ' +
          h +
          '\n        }\n\n        int inIdx = inOffset + ' +
          c +
          ';\n        if (' +
          (1 === l) +
          ') {\n          ' +
          f +
          ' values = ' +
          f +
          '(\n            getValue(batch, inIdx),\n            initializationValue,\n            initializationValue,\n            initializationValue\n          );\n\n          ' +
          h +
          '\n        } else if (' +
          (2 === l) +
          ') {\n          ' +
          f +
          ' values = ' +
          f +
          '(\n            getValue(batch, inIdx),\n            getValue(batch, inIdx + 1),\n            initializationValue,\n            initializationValue\n          );\n\n          ' +
          h +
          '\n        } else if (' +
          (3 === l) +
          ') {\n          ' +
          f +
          ' values = ' +
          f +
          '(\n            getValue(batch, inIdx),\n            getValue(batch, inIdx + 1),\n            getValue(batch, inIdx + 2),\n            initializationValue\n          );\n\n          ' +
          h +
          '\n        }\n        setOutput(' +
          u +
          ');\n      }\n    ')
    },
    Zi = function (t, e) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = t)
      for (var n = '', r = 0; r < 4; r++) {
        var o = 'thisRC = rc;'
        r % 2 == 1 && (o += 'thisRC.z += 1;'),
          r > 1 && (o += 'thisRC.y += 1;'),
          (n +=
            '\n        ' +
            o +
            '\n        ' +
            (r > 0 ? 'if(thisRC.y < rows && thisRC.z < cols){' : '') +
            '\n          int flatIndex = getFlatIndex(thisRC);\n\n          ivec3 inputRC = inputCoordsFromReshapedOutCoords(flatIndex);\n          vec2 inputRCInnerDims = vec2(float(inputRC.y),float(inputRC.z));\n\n          result[' +
            r +
            '] =\n            getChannel(getA(inputRC.x, inputRC.y, inputRC.z), inputRCInnerDims);\n        ' +
            (r > 0 ? '}' : '') +
            '\n      ')
      }
      this.userCode =
        '\n      \n    ivec3 inputCoordsFromReshapedOutCoords(int index) {\n      ' +
        ca(['r', 'c', 'd'], e) +
        '\n      return ivec3(r, c, d);\n    }\n  \n      ' +
        la(t) +
        '\n\n      void main() {\n        ivec3 rc = getOutputCoords();\n\n        vec4 result = vec4(0.);\n\n        ivec3 thisRC;\n        int rows = ' +
        t[1] +
        ';\n        int cols = ' +
        t[2] +
        ';\n\n        ' +
        n +
        '\n\n        setOutput(result);\n      }\n    '
    }
  var ts = function (t, e, n) {
      ;(this.variableNames = ['dy']),
        (this.outputShape = []),
        (this.outputShape = e.shape)
      var r = e.shape,
        o = r[1],
        a = r[2],
        i = t.shape,
        s = i[1],
        u = i[2],
        c = [n && s > 1 ? o - 1 : o, n && u > 1 ? a - 1 : a],
        l = [n && s > 1 ? s - 1 : s, n && u > 1 ? u - 1 : u],
        h = c[0] / l[0],
        f = c[1] / l[1],
        d = 1 / h,
        p = 1 / f,
        v = 2 * Math.ceil(d) + 2,
        m = 2 * Math.ceil(p) + 2
      this.userCode =
        '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n        int r = coords[1];\n        int c = coords[2];\n\n        float accumulator = 0.0;\n\n        const float heightScale = float(' +
        h +
        ');\n        const float widthScale = float(' +
        f +
        ');\n\n        const float invHeightScale = float(' +
        d +
        ');\n        const float invWidthScale = float(' +
        p +
        ');\n\n        const int winHeight = int(' +
        v +
        ');\n        const int winWidth = int(' +
        m +
        ');\n\n        // Compute bounds for where in dy we will look\n        float startRLerp = floor(float(r) * invHeightScale);\n        int startDyR = int(startRLerp - float(winHeight / 2));\n\n        float startCLerp = floor(float(c) * invWidthScale);\n        int startDyC = int(startCLerp - float(winWidth / 2));\n\n        // Loop over dy\n        for (int dyROffset = 0; dyROffset < winHeight; dyROffset++) {\n          int dyR = dyROffset + startDyR;\n\n          // Guard against the window exceeding the bounds of dy\n          if (dyR < 0 || dyR >= ' +
        s +
        ') {\n            continue;\n          }\n\n          for (int dyCOffset = 0; dyCOffset < winWidth; dyCOffset++) {\n            int dyC = dyCOffset + startDyC;\n\n            // Guard against the window exceeding the bounds of dy\n            if (dyC < 0 || dyC >= ' +
        u +
        ') {\n              continue;\n            }\n\n            float dxR = float(dyR) * heightScale;\n            int topDxRIndex = int(floor(dxR));\n            int bottomDxRIndex = int(min(ceil(dxR), ' +
        (o - 1) +
        '.0));\n            float dxRLerp = dxR - float(topDxRIndex);\n            float inverseDxRLerp = 1.0 - dxRLerp;\n\n            float dxC = float(dyC) * widthScale;\n            int leftDxCIndex = int(floor(dxC));\n            int rightDxCIndex = int(min(ceil(dxC), ' +
        (a - 1) +
        '.0));\n            float dxCLerp = dxC - float(leftDxCIndex);\n            float inverseDxCLerp = 1.0 - dxCLerp;\n\n            if (r == topDxRIndex && c == leftDxCIndex) {\n              // topLeft\n              accumulator +=\n                getDy(b, dyR, dyC, d) * inverseDxRLerp * inverseDxCLerp;\n            }\n\n            if (r == topDxRIndex && c == rightDxCIndex) {\n              // topRight\n              accumulator += getDy(b, dyR, dyC, d) * inverseDxRLerp * dxCLerp;\n            }\n\n            if (r == bottomDxRIndex && c == leftDxCIndex) {\n              // bottomLeft\n              accumulator += getDy(b, dyR, dyC, d) * dxRLerp * inverseDxCLerp;\n            }\n\n            if (r == bottomDxRIndex && c == rightDxCIndex) {\n              // bottomRight\n              accumulator += getDy(b, dyR, dyC, d) * dxRLerp * dxCLerp;\n            }\n          }\n        }\n        // End loop over dy\n\n        setOutput(accumulator);\n      }\n    '
    },
    es = function (t, e, n, r) {
      ;(this.variableNames = ['A']), (this.outputShape = [])
      var o = t[0],
        a = t[1],
        i = t[2],
        s = t[3]
      this.outputShape = [o, e, n, s]
      var u = [r && e > 1 ? a - 1 : a, r && n > 1 ? i - 1 : i],
        c = [r && e > 1 ? e - 1 : e, r && n > 1 ? n - 1 : n]
      this.userCode =
        '\n      const vec2 effectiveInputOverOutputRatioRC = vec2(\n          ' +
        u[0] / c[0] +
        ',\n          ' +
        u[1] / c[1] +
        ');\n      const vec2 inputShapeRC = vec2(' +
        a +
        '.0, ' +
        i +
        '.0);\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n        ivec2 yRC = coords.yz;\n\n        // Fractional source index.\n        vec2 sourceFracIndexRC = vec2(yRC) * effectiveInputOverOutputRatioRC;\n\n        // Compute the four integer indices.\n        ivec2 sourceFloorRC = ivec2(sourceFracIndexRC);\n        ivec2 sourceCeilRC = ivec2(\n          min(inputShapeRC - 1.0, ceil(sourceFracIndexRC)));\n\n        float topLeft = getA(b, sourceFloorRC.x, sourceFloorRC.y, d);\n        float bottomLeft = getA(b, sourceCeilRC.x, sourceFloorRC.y, d);\n        float topRight = getA(b, sourceFloorRC.x, sourceCeilRC.y, d);\n        float bottomRight = getA(b, sourceCeilRC.x, sourceCeilRC.y, d);\n\n        vec2 fracRC = sourceFracIndexRC - vec2(sourceFloorRC);\n\n        float top = topLeft + (topRight - topLeft) * fracRC.y;\n        float bottom = bottomLeft + (bottomRight - bottomLeft) * fracRC.y;\n        float newValue = top + (bottom - top) * fracRC.x;\n\n        setOutput(newValue);\n      }\n    '
    },
    ns = function (t, e, n, r) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = [])
      var o = t[0],
        a = t[1],
        i = t[2],
        s = t[3]
      this.outputShape = [o, e, n, s]
      var u = [r && e > 1 ? a - 1 : a, r && n > 1 ? i - 1 : i],
        c = [r && e > 1 ? e - 1 : e, r && n > 1 ? n - 1 : n]
      this.userCode =
        '\n      const vec3 effectiveInputOverOutputRatioRC = vec3(\n          ' +
        u[0] / c[0] +
        ',\n          ' +
        u[1] / c[1] +
        ',\n          ' +
        u[1] / c[1] +
        ');\n      const vec3 inputShapeRC = vec3(' +
        a +
        '.0, ' +
        i +
        '.0,\n                                     ' +
        i +
        '.0);\n\n      float getAValue(int b, int r, int c, int d) {\n        return getChannel(getA(b, r, c, d), vec2(c, d));\n      }\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n        // Calculate values for next column in yRC.z.\n        ivec3 yRC = coords.yzz + ivec3(0, 0, 1);\n\n        // Fractional source index.\n        vec3 sourceFracIndexRC = vec3(yRC) * effectiveInputOverOutputRatioRC;\n\n        // Compute the four integer indices.\n        ivec3 sourceFloorRC = ivec3(sourceFracIndexRC);\n        ivec3 sourceCeilRC = ivec3(\n          min(inputShapeRC - 1.0, ceil(sourceFracIndexRC)));\n\n        // Should we calculate next column and row elements in 2x2 packed cell.\n        bool hasNextCol = d < ' +
        (s - 1) +
        ';\n        bool hasNextRow = coords.z < ' +
        (n - 1) +
        ';\n\n        // In parallel, construct four corners for all four components in\n        // packed 2x2 cell.\n        vec4 topLeft = vec4(\n          getAValue(b, sourceFloorRC.x, sourceFloorRC.y, d),\n          hasNextCol ? getAValue(b, sourceFloorRC.x, sourceFloorRC.y, d + 1)\n                     : 0.0,\n          hasNextRow ? getAValue(b, sourceFloorRC.x, sourceFloorRC.z, d)\n                     : 0.0,\n          (hasNextRow && hasNextCol) ?\n            getAValue(b, sourceFloorRC.x, sourceFloorRC.z, d + 1) : 0.0);\n\n        vec4 bottomLeft = vec4(\n          getAValue(b, sourceCeilRC.x, sourceFloorRC.y, d),\n          hasNextCol ? getAValue(b, sourceCeilRC.x, sourceFloorRC.y, d + 1)\n                     : 0.0,\n          hasNextRow ? getAValue(b, sourceCeilRC.x, sourceFloorRC.z, d)\n                     : 0.0,\n          (hasNextRow && hasNextCol) ?\n            getAValue(b, sourceCeilRC.x, sourceFloorRC.z, d + 1) : 0.0);\n\n        vec4 topRight = vec4(\n          getAValue(b, sourceFloorRC.x, sourceCeilRC.y, d),\n          hasNextCol ? getAValue(b, sourceFloorRC.x, sourceCeilRC.y, d + 1)\n                     : 0.0,\n          hasNextRow ? getAValue(b, sourceFloorRC.x, sourceCeilRC.z, d)\n                     : 0.0,\n          (hasNextRow && hasNextCol) ?\n            getAValue(b, sourceFloorRC.x, sourceCeilRC.z, d + 1) : 0.0);\n\n        vec4 bottomRight = vec4(\n          getAValue(b, sourceCeilRC.x, sourceCeilRC.y, d),\n          hasNextCol ? getAValue(b, sourceCeilRC.x, sourceCeilRC.y, d + 1)\n                     : 0.0,\n          hasNextRow ? getAValue(b, sourceCeilRC.x, sourceCeilRC.z, d)\n                     : 0.0,\n          (hasNextRow && hasNextCol) ?\n            getAValue(b, sourceCeilRC.x, sourceCeilRC.z, d + 1) : 0.0);\n\n        vec3 fracRC = sourceFracIndexRC - vec3(sourceFloorRC);\n\n        vec4 top = mix(topLeft, topRight, fracRC.yyzz);\n        vec4 bottom = mix(bottomLeft, bottomRight, fracRC.yyzz);\n        vec4 newValue = mix(top, bottom, fracRC.x);\n\n        setOutput(newValue);\n      }\n    '
    },
    rs = function (t, e, n) {
      ;(this.variableNames = ['dy']),
        (this.outputShape = []),
        (this.outputShape = e.shape)
      var r = e.shape,
        o = r[1],
        a = r[2],
        i = t.shape,
        s = i[1],
        u = i[2],
        c = [n && s > 1 ? o - 1 : o, n && u > 1 ? a - 1 : a],
        l = [n && s > 1 ? s - 1 : s, n && u > 1 ? u - 1 : u],
        h = c[0] / l[0],
        f = c[1] / l[1],
        d = 1 / h,
        p = 1 / f,
        v = 2 * Math.ceil(d) + 2,
        m = 2 * Math.ceil(p) + 2
      this.userCode =
        '\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n        int r = coords[1];\n        int c = coords[2];\n\n        float accumulator = 0.0;\n\n        const float heightScale = float(' +
        h +
        ');\n        const float widthScale = float(' +
        f +
        ');\n\n        const float invHeightScale = float(' +
        d +
        ');\n        const float invWidthScale = float(' +
        p +
        ');\n\n        const int winHeight = int(' +
        v +
        ');\n        const int winWidth = int(' +
        m +
        ');\n\n        // Compute bounds for where in dy we will look\n        float startRLerp = floor(float(r) * invHeightScale);\n        int startDyR = int(floor(startRLerp - float(winHeight / 2)));\n\n        float startCLerp = floor(float(c) * invWidthScale);\n        int startDyC = int(floor(startCLerp - float(winWidth / 2)));\n\n        // Loop over dy\n        for (int dyROffset = 0; dyROffset < winHeight; dyROffset++) {\n          int dyR = dyROffset + startDyR;\n\n          // Guard against the window exceeding the bounds of dy\n          if (dyR < 0 || dyR >= ' +
        s +
        ') {\n            continue;\n          }\n\n          for (int dyCOffset = 0; dyCOffset < winWidth; dyCOffset++) {\n            int dyC = dyCOffset + startDyC;\n\n            // Guard against the window exceeding the bounds of dy\n            if (dyC < 0 || dyC >= ' +
        u +
        ') {\n              continue;\n            }\n\n            float sourceFracRow =\n              float(' +
        c[0] +
        ') *\n                (float(dyR) / float(' +
        l[0] +
        '));\n\n            float sourceFracCol =\n                float(' +
        c[1] +
        ') *\n                  (float(dyC) / float(' +
        l[1] +
        '));\n\n            int sourceNearestRow = int(min(\n                float(int(' +
        o +
        ') - 1),\n                ' +
        n +
        ' ? float(round(sourceFracRow)) :\n                                  float(floor(sourceFracRow))));\n\n            int sourceNearestCol = int(min(\n                float(int(' +
        a +
        ') - 1),\n                ' +
        n +
        ' ? float(round(sourceFracCol)) :\n                                  float(floor(sourceFracCol))));\n\n            if (r == sourceNearestRow && c == sourceNearestCol) {\n              accumulator += getDy(b, dyR, dyC, d);\n            }\n          }\n        }\n        // End loop over dy\n\n        setOutput(accumulator);\n      }\n    '
    },
    os = function (t, e, n, r) {
      ;(this.variableNames = ['A']), (this.outputShape = [])
      var o = t[0],
        a = t[1],
        i = t[2],
        s = t[3]
      this.outputShape = [o, e, n, s]
      var u = [r && e > 1 ? a - 1 : a, r && n > 1 ? i - 1 : i],
        c = [r && e > 1 ? e - 1 : e, r && n > 1 ? n - 1 : n],
        l = r ? '0.5' : '0.0'
      this.userCode =
        '\n      const vec2 effectiveInputOverOutputRatioRC = vec2(\n          ' +
        u[0] / c[0] +
        ',\n          ' +
        u[1] / c[1] +
        ');\n      const vec2 inputShapeRC = vec2(' +
        a +
        '.0, ' +
        i +
        '.0);\n\n      void main() {\n        ivec4 coords = getOutputCoords();\n        int b = coords[0];\n        int d = coords[3];\n        ivec2 yRC = coords.yz;\n\n        // Fractional source index.\n        vec2 sourceFracIndexRC = vec2(yRC) * effectiveInputOverOutputRatioRC;\n\n        // Compute the coordinators of nearest neighbor point.\n        ivec2 sourceNearestRC = ivec2(\n          min(inputShapeRC - 1.0, floor(sourceFracIndexRC + ' +
        l +
        ')));\n\n        float newValue = getA(b, sourceNearestRC.x, sourceNearestRC.y, d);\n\n        setOutput(newValue);\n      }\n    '
    },
    as = function (t, e) {
      this.variableNames = ['x']
      var n = t.length
      if (n > 4)
        throw new Error(
          'WebGL backend: Reverse of rank-' + n + ' tensor is not yet supported'
        )
      if (((this.outputShape = t), 1 !== n)) {
        var r = t
            .map(function (n, r) {
              return (function (n) {
                return -1 !== e.indexOf(n) && 1 !== t[n]
                  ? t[n] + ' - coords[' + n + '] - 1'
                  : 'coords[' + n + ']'
              })(r)
            })
            .join(','),
          o = wa(n)
        this.userCode =
          '\n      void main() {\n        ' +
          o +
          ' coords = getOutputCoords();\n        setOutput(getX(' +
          r +
          '));\n      }\n    '
      } else
        this.userCode =
          '\n        void main() {\n          int coord = getOutputCoords();\n          setOutput(getX(' +
          t[0] +
          ' - coord - 1));\n        }\n      '
    },
    is = function (t, e) {
      ;(this.variableNames = ['x']),
        (this.packedInputs = !0),
        (this.packedOutput = !0)
      var n = t.length
      if (n > 4)
        throw new Error(
          'WebGL backend: Reverse of rank-' + n + ' tensor is not yet supported'
        )
      this.outputShape = t
      var r = sa('rc', n),
        o = r[n - 1] + ' + 1 < ' + this.outputShape[n - 1],
        a = r[n - 2] + ' + 1 < ' + this.outputShape[n - 2],
        i = wa(n)
      function s(n) {
        var r = t.map(function (r, o) {
          return (function (n, r) {
            return -1 !== e.indexOf(n) && 1 !== t[n]
              ? t[n] + ' - ' + r[n] + ' - 1'
              : '' + r[n]
          })(o, n)
        })
        return (
          'getChannel(getX(' +
          r.join(',') +
          '), vec2(' +
          r.slice(-2).join(',') +
          '))'
        )
      }
      this.userCode =
        1 === n
          ? '\n        void main(){\n          int rc = getOutputCoords();\n          vec4 result = vec4(0.);\n          result.r = getChannel(getX(' +
            t[0] +
            ' - rc - 1),\n            ' +
            t[0] +
            ' - rc - 1);\n          if(' +
            o +
            '){\n              result.g = getChannel(getX(' +
            t[0] +
            ' - (rc  + 1) - 1),\n                ' +
            t[0] +
            ' - (rc  + 1) - 1);\n          }\n          setOutput(result);\n        }\n      '
          : '\n        void main() {\n          ' +
            i +
            ' rc = getOutputCoords();\n          vec4 result = vec4(0.);\n          result.r = ' +
            (function (t) {
              return s(t)
            })(r.slice()) +
            ';\n          if(' +
            o +
            '){\n            result.g = ' +
            (function (t) {
              return (t[n - 1] = '(' + t[n - 1] + ' + 1)'), s(t)
            })(r.slice()) +
            ';\n          }\n          if(' +
            a +
            ') {\n            result.b = ' +
            (function (t) {
              return (t[n - 2] = '(' + t[n - 2] + ' + 1)'), s(t)
            })(r.slice()) +
            ';\n            if(' +
            o +
            ') {\n              result.a = ' +
            (function (t) {
              return (
                (t[n - 1] = '(' + t[n - 1] + ' + 1)'),
                (t[n - 2] = '(' + t[n - 2] + ' + 1)'),
                s(t)
              )
            })(r.slice()) +
            ';\n            }\n          }\n          setOutput(result);\n        }\n    '
    },
    ss = function (t, e, n, r, o, a, i) {
      ;(this.variableNames = ['updates', 'indices', 'defaultValue']),
        (this.outputShape = a)
      var s = wa(o.length),
        u = wa(a.length),
        c = ''
      1 === n ? (c = 'i') : 2 === n && (c = 'i, j')
      var l = 'getIndices(' + c + ')',
        h = ''
      1 === r ? (h = 'i') : 2 === r && (h = 'i, coords[1]')
      var f = 'getUpdates(' + h + ')',
        d = e > 1 ? 'strides[j]' : 'strides'
      this.userCode =
        '\n        ' +
        s +
        ' strides = ' +
        s +
        '(' +
        o +
        ');\n\n        void main() {\n          ' +
        u +
        ' coords = getOutputCoords();\n          float sum = 0.0;\n          bool found = false;\n          for (int i = 0; i < ' +
        t +
        '; i++) {\n            int flattenedIndex = 0;\n            for (int j = 0; j < ' +
        e +
        '; j++) {\n              int index = round(' +
        l +
        ');\n              flattenedIndex += index * ' +
        d +
        ';\n            }\n            if (flattenedIndex == coords[0]) {\n              sum += ' +
        f +
        ';\n              found = true;\n            }\n          }\n          setOutput(mix(getDefaultValue(), sum, float(found)));\n        }\n      '
    },
    us = function (t, e) {
      this.variableNames = ['x', 'segmentIds']
      var n = t.windowSize,
        r = t.batchSize,
        o = t.inSize,
        a = t.numSegments,
        i = a * Math.ceil(o / n)
      this.outputShape = [r, i]
      var s = 4 * Math.floor(n / 4),
        u = n % 4,
        c = '\n        sumValue += dot(values, segFilter);\n    ',
        l = ''
      o % n > 0 &&
        (l =
          '\n        if (inIdx < 0 || inIdx >= ' +
          o +
          ') {\n          return initializationValue;\n        }\n      ')
      var h = ''
      o % n > 0 &&
        (h =
          '\n        if (inIdx < 0 || inIdx >= ' +
          o +
          ') {\n          return -1.0;\n        }\n      '),
        (this.userCode =
          '\n      const float initializationValue = 0.0;\n\n      float getValue(int batch, int inIdx) {\n        ' +
          l +
          '\n        return getX(batch, inIdx);\n      }\n\n      float getSegmentIdAtIndex(int inIdx) {\n        ' +
          h +
          '\n        return getSegmentIds(inIdx);\n      }\n\n      void main() {\n        ivec2 coords = getOutputCoords();\n        int batch = coords[0];\n        int outIdx = coords[1];\n        int inOffset = int(floor(float(outIdx) / float(\n          ' +
          a +
          ')) * float(' +
          n +
          '));\n        int currentSeg = int(mod(float(outIdx), float(' +
          a +
          ')));\n\n        float sumValue = 0.0;\n\n        for (int i = 0; i < ' +
          s +
          '; i += 4) {\n          int inIdx = inOffset + i;\n          vec4 values = vec4(\n            getValue(batch, inIdx),\n            getValue(batch, inIdx + 1),\n            getValue(batch, inIdx + 2),\n            getValue(batch, inIdx + 3)\n          );\n\n          vec4 segFilter = vec4(\n            int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,\n            int(getSegmentIdAtIndex(inIdx + 1)) == currentSeg ? 1 : 0,\n            int(getSegmentIdAtIndex(inIdx + 2)) == currentSeg ? 1 : 0,\n            int(getSegmentIdAtIndex(inIdx + 3)) == currentSeg ? 1 : 0\n          );\n\n          ' +
          c +
          '\n        }\n\n        int inIdx = inOffset + ' +
          s +
          ';\n        if (' +
          (1 === u) +
          ') {\n          vec4 values = vec4(\n            getValue(batch, inIdx),\n            initializationValue,\n            initializationValue,\n            initializationValue\n          );\n\n          int inIdxSeg = int(getSegmentIdAtIndex(inIdx));\n\n          vec4 segFilter = vec4(\n            int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,\n            0,\n            0,\n            0\n          );\n\n          ' +
          c +
          '\n        } else if (' +
          (2 === u) +
          ') {\n          vec4 values = vec4(\n            getValue(batch, inIdx),\n            getValue(batch, inIdx + 1),\n            initializationValue,\n            initializationValue\n          );\n\n          vec4 segFilter = vec4(\n            int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,\n            int(getSegmentIdAtIndex(inIdx + 1)) == currentSeg ? 1 : 0,\n              0,\n              0\n          );\n\n          ' +
          c +
          '\n        } else if (' +
          (3 === u) +
          ') {\n          vec4 values = vec4(\n            getValue(batch, inIdx),\n            getValue(batch, inIdx + 1),\n            getValue(batch, inIdx + 2),\n            initializationValue\n          );\n\n          vec4 segFilter = vec4(\n            int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,\n            int(getSegmentIdAtIndex(inIdx + 1)) == currentSeg ? 1 : 0,\n            int(getSegmentIdAtIndex(inIdx + 2)) == currentSeg ? 1 : 0,\n            0\n          );\n\n          ' +
          c +
          '\n        }\n        setOutput(sumValue);\n      }\n    ')
    },
    cs = function (t, e, n) {
      var r, o
      if (
        ((this.variableNames = ['c', 'a', 'b']), (this.outputShape = e), n > 4)
      )
        throw Error('Where for rank ' + n + ' is not yet supported')
      if (1 === n) (o = 'resRC'), (r = 'resRC')
      else {
        for (
          var a = ['resRC.x', 'resRC.y', 'resRC.z', 'resRC.w'],
            i = [],
            s = [],
            u = 0;
          u < e.length;
          u++
        )
          s.push('' + a[u]), u < t && i.push('' + a[u])
        ;(r = i.join()), (o = s.join())
      }
      var c = wa(n)
      this.userCode =
        '\n      void main() {\n        ' +
        c +
        ' resRC = getOutputCoords();\n        float cVal = getC(' +
        r +
        ');\n        if (cVal >= 1.0) {\n          setOutput(getA(' +
        o +
        '));\n        } else {\n          setOutput(getB(' +
        o +
        '));\n        }\n      }\n    '
    },
    ls = (function () {
      function t(t) {
        ;(this.variableNames = ['source']),
          (this.outputShape = t),
          (this.rank = t.length)
        var e,
          n = wa(this.rank),
          r = 'uniform int start[' + this.rank + '];',
          o = (function (t) {
            if (1 === t) return 'sourceLoc'
            if (t <= 6)
              return hs
                .slice(0, t)
                .map(function (t) {
                  return 'sourceLoc.' + t
                })
                .join(',')
            throw Error('Slicing for rank ' + t + ' is not yet supported')
          })(this.rank)
        ;(e =
          '\n        ' +
          n +
          ' sourceLoc;\n        ' +
          n +
          ' coords = getOutputCoords();\n        ' +
          t
            .map(function (t, e) {
              return (
                'sourceLoc.' +
                hs[e] +
                ' = start[' +
                e +
                '] + coords.' +
                hs[e] +
                ';'
              )
            })
            .join('\n') +
          '\n      '),
          (this.userCode =
            '\n      ' +
            r +
            '\n      void main() {\n        ' +
            e +
            '\n        setOutput(getSource(' +
            o +
            '));\n      }\n    ')
      }
      return (
        (t.prototype.getCustomSetupFunc = function (t) {
          var e = this
          if (t.length !== this.rank)
            throw Error(
              'The rank (' +
                this.rank +
                ') of the program must match the length of start (' +
                t.length +
                ')'
            )
          return function (n, r) {
            ;(null == e.startLoc &&
              ((e.startLoc = n.getUniformLocationNoThrow(r, 'start')),
              null == e.startLoc)) ||
              n.gl.uniform1iv(e.startLoc, t)
          }
        }),
        t
      )
    })(),
    hs = ['x', 'y', 'z', 'w', 'u', 'v']
  var fs = (function () {
      function t(t) {
        ;(this.variableNames = ['source']),
          (this.packedInputs = !0),
          (this.packedOutput = !0),
          (this.outputShape = t),
          (this.rank = t.length)
        var e = wa(this.rank),
          n = sa('coords', this.rank),
          r = sa('sourceLoc', this.rank),
          o =
            1 === this.rank ? 'sourceLoc' : 'vec2(' + r.slice(-2).join() + ')',
          a = 'getChannel(getSource(' + r.join() + '), ' + o + ')',
          i =
            '\n      result.x = ' +
            a +
            ';\n      if (++' +
            n[this.rank - 1] +
            ' < ' +
            t[this.rank - 1] +
            ') {\n        ++' +
            r[this.rank - 1] +
            ';\n        result.y = ' +
            a +
            ';\n        --' +
            r[this.rank - 1] +
            ';\n      }\n    ',
          s =
            1 === this.rank
              ? ''
              : '\n      --' +
                n[this.rank - 1] +
                ';\n      if (++' +
                n[this.rank - 2] +
                ' < ' +
                t[this.rank - 2] +
                ') {\n        ++' +
                r[this.rank - 2] +
                ';\n        result.z = ' +
                a +
                ';\n        if (++' +
                n[this.rank - 1] +
                ' < ' +
                t[this.rank - 1] +
                ') {\n          ++' +
                r[this.rank - 1] +
                ';\n          result.w = ' +
                a +
                ';\n        }\n      }\n    ',
          u =
            this.rank <= 4
              ? 'sourceLoc = coords +\n            ' +
                e +
                '(' +
                t
                  .map(function (t, e) {
                    return 'start[' + e + ']'
                  })
                  .join() +
                ');'
              : t
                  .map(function (t, e) {
                    return r[e] + ' = ' + n[e] + ' + start[' + e + '];'
                  })
                  .join('\n')
        this.userCode =
          '\n      uniform int start[' +
          this.rank +
          '];\n      void main() {\n        ' +
          e +
          ' coords = getOutputCoords();\n        ' +
          e +
          ' sourceLoc;\n        ' +
          u +
          '\n        vec4 result = vec4(0.);\n        ' +
          i +
          '\n        ' +
          s +
          '\n        setOutput(result);\n      }\n    '
      }
      return (
        (t.prototype.getCustomSetupFunc = function (t) {
          var e = this
          if (t.length !== this.rank)
            throw Error(
              'The rank (' +
                this.rank +
                ') of the program must match the length of start (' +
                t.length +
                ')'
            )
          return function (n, r) {
            ;(null == e.startLoc &&
              ((e.startLoc = n.getUniformLocationNoThrow(r, 'start')),
              null == e.startLoc)) ||
              n.gl.uniform1iv(e.startLoc, t)
          }
        }),
        t
      )
    })(),
    ds = function (t, e, n) {
      ;(this.variableNames = ['x']), (this.outputShape = n)
      var r = n.length,
        o = wa(n.length),
        a = wa(n.length),
        i = ''
      if (1 === r) i = 'coords * strides + begin'
      else {
        var s = 0
        i = n
          .map(function (t, e) {
            return (
              s++,
              1 === n.length
                ? 'coords * strides[' + e + '] + begin[' + e + ']'
                : 'coords[' +
                  (s - 1) +
                  '] * strides[' +
                  e +
                  '] + begin[' +
                  e +
                  ']'
            )
          })
          .join(',')
      }
      this.userCode =
        '\n      ' +
        o +
        ' begin = ' +
        o +
        '(' +
        t +
        ');\n      ' +
        o +
        ' strides = ' +
        o +
        '(' +
        e +
        ');\n\n      void main() {\n        ' +
        a +
        ' coords = getOutputCoords();\n        setOutput(getX(' +
        i +
        '));\n      }\n    '
    },
    ps = (function () {
      function t(t) {
        ;(this.gpgpu = t),
          (this.numUsedTextures = 0),
          (this.numFreeTextures = 0),
          (this.freeTextures = {}),
          (this.logEnabled = !1),
          (this.usedTextures = {})
      }
      return (
        (t.prototype.acquireTexture = function (t, e, n) {
          var r,
            o = vs(e, n),
            a = ms(t, o, n)
          if (
            (a in this.freeTextures || (this.freeTextures[a] = []),
            a in this.usedTextures || (this.usedTextures[a] = []),
            this.freeTextures[a].length > 0)
          ) {
            this.numFreeTextures--, this.numUsedTextures++, this.log()
            var i = this.freeTextures[a].shift()
            return this.usedTextures[a].push(i), i
          }
          return (
            this.numUsedTextures++,
            this.log(),
            o === Gt.PACKED_2X2_FLOAT32
              ? (r = this.gpgpu.createPackedMatrixTexture(t[0], t[1]))
              : o === Gt.PACKED_2X2_FLOAT16
              ? (r = this.gpgpu.createFloat16PackedMatrixTexture(t[0], t[1]))
              : o === Gt.UNPACKED_FLOAT32
              ? (r = this.gpgpu.createFloat32MatrixTexture(t[0], t[1]))
              : o === Gt.UNPACKED_FLOAT16
              ? (r = this.gpgpu.createFloat16MatrixTexture(t[0], t[1]))
              : o === Gt.PACKED_4X1_UNSIGNED_BYTE &&
                (r = this.gpgpu.createUnsignedBytesMatrixTexture(t[0], t[1])),
            this.usedTextures[a].push(r),
            r
          )
        }),
        (t.prototype.releaseTexture = function (t, e, n, r) {
          if (null != this.freeTextures) {
            var o = ms(e, vs(n, r), r)
            o in this.freeTextures || (this.freeTextures[o] = []),
              this.freeTextures[o].push(t),
              this.numFreeTextures++,
              this.numUsedTextures--
            var a = this.usedTextures[o],
              i = a.indexOf(t)
            if (i < 0)
              throw new Error(
                'Cannot release a texture that was never provided by this texture manager'
              )
            a.splice(i, 1), this.log()
          }
        }),
        (t.prototype.log = function () {
          if (this.logEnabled) {
            var t = this.numFreeTextures + this.numUsedTextures
            console.log(
              'Free/Used',
              this.numFreeTextures + ' / ' + this.numUsedTextures,
              '(' + t + ')'
            )
          }
        }),
        (t.prototype.getNumUsedTextures = function () {
          return this.numUsedTextures
        }),
        (t.prototype.getNumFreeTextures = function () {
          return this.numFreeTextures
        }),
        (t.prototype.dispose = function () {
          var t = this
          if (null != this.freeTextures) {
            for (var e in this.freeTextures)
              this.freeTextures[e].forEach(function (e) {
                t.gpgpu.deleteMatrixTexture(e)
              })
            for (var e in this.usedTextures)
              this.usedTextures[e].forEach(function (e) {
                t.gpgpu.deleteMatrixTexture(e)
              })
            ;(this.freeTextures = null),
              (this.usedTextures = null),
              (this.numUsedTextures = 0),
              (this.numFreeTextures = 0)
          }
        }),
        t
      )
    })()
  function vs(t, e) {
    if (t === zt.UPLOAD) return Gt.PACKED_2X2_FLOAT32
    if (t === zt.RENDER || null == t)
      return (function (t) {
        return i().getBool('WEBGL_RENDER_FLOAT32_ENABLED')
          ? t
            ? Gt.PACKED_2X2_FLOAT32
            : Gt.UNPACKED_FLOAT32
          : t
          ? Gt.PACKED_2X2_FLOAT16
          : Gt.UNPACKED_FLOAT16
      })(e)
    if (t === zt.DOWNLOAD || t === zt.PIXELS) return Gt.PACKED_4X1_UNSIGNED_BYTE
    throw new Error('Unknown logical texture type ' + t)
  }
  function ms(t, e, n) {
    return t[0] + '_' + t[1] + '_' + e + '_' + n
  }
  var gs = function (t, e) {
    this.variableNames = ['A']
    for (var n = new Array(t.length), r = 0; r < n.length; r++)
      n[r] = t[r] * e[r]
    ;(this.outputShape = n), (this.rank = n.length)
    var o = wa(this.rank),
      a = (function (t) {
        var e = t.length
        if (e > 5) throw Error('Tile for rank ' + e + ' is not yet supported')
        if (1 === e) return 'imod(resRC, ' + t[0] + ')'
        for (
          var n = ['resRC.x', 'resRC.y', 'resRC.z', 'resRC.w', 'resRC.u'],
            r = [],
            o = 0;
          o < t.length;
          o++
        )
          r.push('imod(' + n[o] + ', ' + t[o] + ')')
        return r.join()
      })(t)
    this.userCode =
      '\n      void main() {\n        ' +
      o +
      ' resRC = getOutputCoords();\n        setOutput(getA(' +
      a +
      '));\n      }\n    '
  }
  var ys = function (t, e) {
    this.variableNames = ['A']
    for (var n = new Array(t.length), r = 0; r < n.length; r++) n[r] = t[e[r]]
    ;(this.outputShape = n), (this.rank = n.length)
    var o = wa(this.rank),
      a = (function (t) {
        var e = t.length
        if (e > 6)
          throw Error('Transpose for rank ' + e + ' is not yet supported')
        for (
          var n = [
              'resRC.x',
              'resRC.y',
              'resRC.z',
              'resRC.w',
              'resRC.u',
              'resRC.v',
            ],
            r = new Array(e),
            o = 0;
          o < t.length;
          o++
        )
          r[t[o]] = n[o]
        return r.join()
      })(e)
    this.userCode =
      '\n    void main() {\n      ' +
      o +
      ' resRC = getOutputCoords();\n      setOutput(getA(' +
      a +
      '));\n    }\n    '
  }
  var xs = function (t, e) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0)
      for (var n = new Array(t.length), r = 0; r < n.length; r++) n[r] = t[e[r]]
      if (((this.outputShape = n), (this.rank = n.length), this.rank > 6))
        throw Error(
          'Packed transpose for rank ' + this.rank + ' is not yet supported.'
        )
      var o = wa(this.rank),
        a = ia('rc', this.rank),
        i = new Array(this.rank)
      for (r = 0; r < e.length; r++) i[e[r]] = a[r]
      var s = 'vec2(' + i.slice(-2).join() + ')',
        u = '++' + a[this.rank - 1] + ' < ' + n[this.rank - 1],
        c = 'getChannel(getA(' + i.join() + '), ' + s + ')'
      this.userCode =
        '\n    void main() {\n      ' +
        o +
        ' rc = getOutputCoords();\n      vec4 result = vec4(0.);\n      result[0] = ' +
        c +
        ';\n      if(' +
        u +
        ') {\n        result[1] = ' +
        c +
        ';\n      }\n      --' +
        a[this.rank - 1] +
        ';\n      if(++' +
        a[this.rank - 2] +
        ' < ' +
        n[this.rank - 2] +
        ') {\n        result[2] = ' +
        c +
        ';\n        if(' +
        u +
        ') {\n          result[3] = ' +
        c +
        ';\n        }\n      }\n      setOutput(result);\n    }\n    '
    },
    bs = 1.7580993408473768,
    ws = 1.0507009873554805,
    Cs = function (t, e) {
      ;(this.variableNames = ['A']),
        (this.outputShape = t),
        (this.userCode =
          '\n      float unaryOperation(float x) {\n        ' +
          e +
          '\n      }\n\n      void main() {\n        float x = getAAtOutCoords();\n        float y = unaryOperation(x);\n\n        setOutput(y);\n      }\n    ')
    },
    Es = 'if (isnan(x)) return x;',
    Rs = 'return x;',
    Is = 'return abs(x);',
    ks = Es + '\n  return (x < 0.0) ? 0.0 : x;\n',
    Ss = Es + '\n  return (x < 0.0) ? 0.0 : min(6.0, x);\n',
    As = 'return (x >= 0.0) ? x : (exp(x) - 1.0);',
    Ds =
      '\n  // Stable and Attracting Fixed Point (0, 1) for Normalized Weights.\n  // see: https://arxiv.org/abs/1706.02515\n  float scaleAlpha = ' +
      bs +
      ';\n  float scale = ' +
      ws +
      ';\n  return (x >= 0.0) ? scale * x : scaleAlpha * (exp(x) - 1.0);\n'
  var Ts = 'return -x;',
    Ns = 'return ceil(x);',
    Fs = 'return floor(x);',
    _s = 'return exp(x);',
    Os = 'return exp(x) - 1.0;',
    Ms = Es + '\n  return sin(x);\n',
    Bs = Es + '\n  return cos(x);\n',
    Ps =
      Es + '\n  if (abs(x) > 1.) {\n    return NAN;\n  }\n  return asin(x);\n',
    Ls =
      Es + '\n  if (abs(x) > 1.) {\n    return NAN;\n  }\n  return acos(x);\n',
    Ws = Es + '\n  return atan(x);\n',
    Us = Es + 'return log(x + sqrt(x * x + 1.0));',
    Vs =
      Es + '\n  if (x < 1.0) return NAN;\n  return log(x + sqrt(x * x - 1.0));',
    zs =
      Es +
      '\n  if ((x < -1.0) || (x > 1.0)) return NAN;\n  return (log(1.0 + x) - log(1.0 - x)) / 2.0;',
    Gs = 'return x;',
    Hs = 'return x;',
    qs =
      '\n  vec4 result = x * vec4(greaterThanEqual(x, vec4(0.0)));\n  bvec4 isNaN = isnan(x);\n\n  result.r = isNaN.r ? x.r : result.r;\n  result.g = isNaN.g ? x.g : result.g;\n  result.b = isNaN.b ? x.b : result.b;\n  result.a = isNaN.a ? x.a : result.a;\n\n  return result;\n',
    Ks =
      '\n  vec4 result = min(x, vec4(6.)) * vec4(greaterThanEqual(x, vec4(0.0)));\n  bvec4 isNaN = isnan(x);\n\n  result.r = isNaN.r ? x.r : result.r;\n  result.g = isNaN.g ? x.g : result.g;\n  result.b = isNaN.b ? x.b : result.b;\n  result.a = isNaN.a ? x.a : result.a;\n\n  return result;\n',
    js =
      '\n  vec4 result;\n\n  result.r = (x.r >= 0.0) ? x.r : (exp(x.r) - 1.0);\n  result.g = (x.g >= 0.0) ? x.g : (exp(x.g) - 1.0);\n  result.b = (x.b >= 0.0) ? x.b : (exp(x.b) - 1.0);\n  result.a = (x.a >= 0.0) ? x.a : (exp(x.a) - 1.0);\n\n  return result;\n',
    Xs = function (t, e) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !0),
        (this.outputShape = t),
        (this.userCode =
          '\n      vec4 unaryOperation(vec4 x) {\n        ' +
          e +
          '\n      }\n\n      void main() {\n        vec4 x = getAAtOutCoords();\n        vec4 y = unaryOperation(x);\n\n        setOutput(y);\n      }\n    ')
    },
    Ys = function (t) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !0),
        (this.packedOutput = !1),
        (this.outputShape = t)
      var e = t.length,
        n = sa('rc', e),
        r = wa(e),
        o = (function (t, e) {
          if (1 === t) return 'rc'
          for (var n = '', r = 0; r < t; r++)
            (n += e[r]), r < t - 1 && (n += ',')
          return n
        })(e, n),
        a = n.slice(-2),
        i = e <= 1 ? 'rc' : 'vec2(' + a.join(',') + ')'
      this.userCode =
        '\n      void main() {\n        ' +
        r +
        ' rc = getOutputCoords();\n        vec4 packedInput = getA(' +
        o +
        ');\n\n        setOutput(getChannel(packedInput, ' +
        i +
        '));\n      }\n    '
    },
    $s = {}
  function Qs(t, e) {
    if ((void 0 === e && (e = !1), 'linear' === t)) return e ? Hs : Rs
    if ('relu' === t) return e ? qs : ks
    if ('elu' === t) return e ? js : As
    if ('relu6' === t) return e ? Ks : Ss
    if ('prelu' === t) return e ? Pa : Ma
    throw new Error(
      'Activation ' + t + ' has not been implemented for the WebGL backend.'
    )
  }
  var Js = 600
  var Zs = (function (t) {
    function o(e) {
      var n,
        r = t.call(this) || this
      if (
        ((r.pendingRead = new WeakMap()),
        (r.pendingDisposal = new WeakSet()),
        (r.dataRefCount = new WeakMap()),
        (r.numBytesInGPU = 0),
        (r.uploadWaitMs = 0),
        (r.downloadWaitMs = 0),
        (r.warnedAboutMemory = !1),
        (r.pendingDeletes = 0),
        (r.disposed = !1),
        !i().getBool('HAS_WEBGL'))
      )
        throw new Error('WebGL is not supported on this device')
      if (null == e) {
        var o = jt(i().getNumber('WEBGL_VERSION'))
        ;(r.binaryCache =
          (n = i().getNumber('WEBGL_VERSION')) in $s
            ? $s[n]
            : (($s[n] = {}), $s[n])),
          (r.gpgpu = new Bi(o)),
          (r.canvas = o.canvas),
          (r.gpgpuCreatedLocally = !0)
      } else
        (r.gpgpu = e),
          (r.binaryCache = {}),
          (r.gpgpuCreatedLocally = !1),
          (r.canvas = e.gl.canvas)
      return (
        (r.textureManager = new ps(r.gpgpu)),
        (r.numMBBeforeWarning =
          null == i().global.screen
            ? 1024
            : (i().global.screen.height *
                i().global.screen.width *
                window.devicePixelRatio *
                Js) /
              1024 /
              1024),
        (r.texData = new xo(r, Lt)),
        r
      )
    }
    return (
      e(o, t),
      (o.prototype.numDataIds = function () {
        return (
          this.texData.numDataIds() +
          (this.cpuBackend ? this.cpuBackend.numDataIds() : 0) -
          this.pendingDeletes
        )
      }),
      (o.prototype.write = function (t, e, n) {
        if (
          (i().getBool('DEBUG') && this.checkNumericalProblems(t),
          'complex64' === n && null != t)
        )
          throw new Error(
            'Cannot write to a complex64 dtype. Please use tf.complex(real, imag).'
          )
        var r = {}
        return (
          this.texData.set(r, {
            shape: e,
            dtype: n,
            values: t,
            usage: zt.UPLOAD,
          }),
          r
        )
      }),
      (o.prototype.move = function (t, e, n, r) {
        if (
          (i().getBool('DEBUG') && this.checkNumericalProblems(e),
          'complex64' === r)
        )
          throw new Error(
            'Cannot write to a complex64 dtype. Please use tf.complex(real, imag).'
          )
        this.texData.set(t, { shape: n, dtype: r, values: e, usage: zt.UPLOAD })
      }),
      (o.prototype.readSync = function (t) {
        var e = this.texData.get(t),
          n = e.values,
          r = e.dtype,
          o = e.complexTensors,
          a = e.slice,
          i = e.shape,
          s = e.isPacked
        if (null != a) {
          var u = void 0
          u = s ? new Xs(i, Gs) : new Cs(i, Gs)
          var c = this.runWebGLProgram(
              u,
              [{ dataId: t, shape: i, dtype: r }],
              r
            ),
            l = this.readSync(c.dataId)
          return this.disposeData(c.dataId), l
        }
        if (null != n) return this.convertAndCacheOnCPU(t)
        if ('string' === r) return n
        var h,
          f,
          d = null != this.activeTimers
        ;(d && (h = et()), 'complex64' === r)
          ? (f = Vo(o.real.dataSync(), o.imag.dataSync()))
          : (f = this.getValuesFromTexture(t))
        return (
          d && (this.downloadWaitMs += et() - h),
          this.convertAndCacheOnCPU(t, f)
        )
      }),
      (o.prototype.read = function (t) {
        return n(this, void 0, void 0, function () {
          var e, n, o, a, s, u, c, l, h, f, d, p, v, m, g, y, x, b, w, C, E, R
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                if (this.pendingRead.has(t))
                  return (
                    (e = this.pendingRead.get(t)),
                    [
                      2,
                      new Promise(function (t) {
                        return e.push(t)
                      }),
                    ]
                  )
                if (
                  ((n = this.texData.get(t)),
                  (o = n.values),
                  (a = n.shape),
                  (s = n.slice),
                  (u = n.dtype),
                  (c = n.complexTensors),
                  (l = n.isPacked),
                  null != s)
                )
                  return (
                    (h = void 0),
                    (h = l ? new Xs(a, Gs) : new Cs(a, Gs)),
                    (f = this.runWebGLProgram(
                      h,
                      [{ dataId: t, shape: a, dtype: u }],
                      u
                    )),
                    (d = this.read(f.dataId)),
                    this.disposeData(f.dataId),
                    [2, d]
                  )
                if (null != o) return [2, this.convertAndCacheOnCPU(t)]
                if (
                  !i().getBool('WEBGL_DOWNLOAD_FLOAT_ENABLED') &&
                  2 === i().getNumber('WEBGL_VERSION')
                )
                  throw new Error(
                    'tensor.data() with WEBGL_DOWNLOAD_FLOAT_ENABLED=false and WEBGL_VERSION=2 not yet supported.'
                  )
                return (
                  (p = null),
                  'complex64' !== u &&
                    i().get('WEBGL_BUFFER_SUPPORTED') &&
                    ((v = this.decode(t)),
                    (m = this.texData.get(v.dataId)),
                    (p = (R = this.gpgpu).createBufferFromTexture.apply(
                      R,
                      [m.texture].concat(Yt(a))
                    ))),
                  this.pendingRead.set(t, []),
                  'complex64' === u
                    ? [3, 2]
                    : [4, this.gpgpu.createAndWaitForFence()]
                )
              case 1:
                r.sent(), (r.label = 2)
              case 2:
                return 'complex64' !== u
                  ? [3, 4]
                  : [4, Promise.all([c.real.data(), c.imag.data()])]
              case 3:
                return (
                  (y = r.sent()), (x = y[0]), (b = y[1]), (g = Vo(x, b)), [3, 5]
                )
              case 4:
                null == p
                  ? (g = this.getValuesFromTexture(t))
                  : ((w = k(a)),
                    (g = this.gpgpu.downloadFloat32MatrixFromBuffer(p, w))),
                  (r.label = 5)
              case 5:
                return (
                  null != v && this.disposeData(v.dataId),
                  (C = this.convertAndCacheOnCPU(t, g)),
                  (E = this.pendingRead.get(t)),
                  this.pendingRead.delete(t),
                  E.forEach(function (t) {
                    return t(C)
                  }),
                  this.pendingDisposal.has(t) &&
                    (this.pendingDisposal.delete(t),
                    this.disposeData(t),
                    this.pendingDeletes--),
                  [2, C]
                )
            }
          })
        })
      }),
      (o.prototype.checkNumericalProblems = function (t) {
        if (null != t)
          for (var e = 0; e < t.length; e++) {
            var n = t[e]
            if (!ee(n)) {
              if (i().getBool('WEBGL_RENDER_FLOAT32_CAPABLE'))
                throw Error(
                  'The value ' +
                    n +
                    " cannot be represented with your current settings. Consider enabling float32 rendering: 'tf.env().set('WEBGL_RENDER_FLOAT32_ENABLED', true);'"
                )
              throw Error(
                'The value ' + n + ' cannot be represented on this device.'
              )
            }
          }
      }),
      (o.prototype.getValuesFromTexture = function (t) {
        var e,
          n = this.texData.get(t),
          r = n.shape,
          o = n.dtype,
          a = n.isPacked,
          s = k(r)
        if (i().getBool('WEBGL_DOWNLOAD_FLOAT_ENABLED')) {
          var u = this.decode(t),
            c = this.texData.get(u.dataId),
            l = (e = this.gpgpu).downloadMatrixFromPackedTexture
              .apply(e, [c.texture].concat(Yt(r)))
              .subarray(0, s)
          return this.disposeData(u.dataId), l
        }
        var h = i().getBool('WEBGL_PACK') && !0 === a,
          f = h ? Te(r) : r,
          d = h ? new ci(f) : new ui(f),
          p = this.runWebGLProgram(
            d,
            [{ shape: f, dtype: o, dataId: t }],
            'float32'
          ),
          v = this.texData.get(p.dataId),
          m = this.gpgpu
            .downloadByteEncodedFloatMatrixFromOutputTexture(
              v.texture,
              v.texShape[0],
              v.texShape[1]
            )
            .subarray(0, s)
        return this.disposeData(p.dataId), m
      }),
      (o.prototype.time = function (t) {
        return n(this, void 0, void 0, function () {
          var e, n, o, a, s, u, c
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                return (
                  (e = this.activeTimers),
                  (n = []),
                  (o = !1),
                  null == this.programTimersStack
                    ? ((this.programTimersStack = n), (o = !0))
                    : this.activeTimers.push(n),
                  (this.activeTimers = n),
                  t(),
                  (a = I(
                    this.activeTimers.map(function (t) {
                      return t.query
                    })
                  ).filter(function (t) {
                    return null != t
                  })),
                  (s = I(
                    this.activeTimers.map(function (t) {
                      return t.name
                    })
                  ).filter(function (t) {
                    return null != t
                  })),
                  (this.activeTimers = e),
                  o && (this.programTimersStack = null),
                  (u = {
                    uploadWaitMs: this.uploadWaitMs,
                    downloadWaitMs: this.downloadWaitMs,
                    kernelMs: null,
                    wallMs: null,
                  }),
                  i().getNumber(
                    'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE'
                  ) > 0
                    ? [4, Promise.all(a)]
                    : [3, 2]
                )
              case 1:
                return (
                  (c = r.sent()),
                  (u.kernelMs = w(c)),
                  (u.getExtraProfileInfo = function () {
                    return c
                      .map(function (t, e) {
                        return { name: s[e], ms: t }
                      })
                      .map(function (t) {
                        return t.name + ': ' + t.ms
                      })
                      .join(', ')
                  }),
                  [3, 3]
                )
              case 2:
                ;(u.kernelMs = {
                  error:
                    'WebGL query timers are not supported in this environment.',
                }),
                  (r.label = 3)
              case 3:
                return (
                  (this.uploadWaitMs = 0), (this.downloadWaitMs = 0), [2, u]
                )
            }
          })
        })
      }),
      (o.prototype.memory = function () {
        return { unreliable: !1, numBytesInGPU: this.numBytesInGPU }
      }),
      (o.prototype.startTimer = function () {
        return i().getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE') >
          0
          ? this.gpgpu.beginQuery()
          : { startMs: et(), endMs: null }
      }),
      (o.prototype.endTimer = function (t) {
        return i().getNumber('WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE') >
          0
          ? (this.gpgpu.endQuery(), t)
          : ((t.endMs = et()), t)
      }),
      (o.prototype.getQueryTime = function (t) {
        return n(this, void 0, void 0, function () {
          var e
          return r(this, function (n) {
            return i().getNumber(
              'WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE'
            ) > 0
              ? [2, this.gpgpu.waitForQueryAndGetTime(t)]
              : [2, (e = t).endMs - e.startMs]
          })
        })
      }),
      (o.prototype.disposeData = function (t) {
        if (!this.pendingDisposal.has(t)) {
          if (this.pendingRead.has(t))
            return this.pendingDisposal.add(t), void this.pendingDeletes++
          if (this.texData.has(t)) {
            this.releaseGPUData(t)
            var e = this.texData.get(t).complexTensors
            null != e && (e.real.dispose(), e.imag.dispose()),
              this.texData.delete(t)
          }
        }
      }),
      (o.prototype.releaseGPUData = function (t) {
        var e = this.texData.get(t),
          n = e.texture,
          r = e.dtype,
          o = e.texShape,
          a = e.usage,
          i = e.isPacked,
          s = e.slice,
          u = (s && s.origDataId) || t,
          c = this.dataRefCount.get(u)
        c > 1
          ? this.dataRefCount.set(u, c - 1)
          : (this.dataRefCount.delete(u),
            null != n &&
              ((this.numBytesInGPU -= this.computeBytes(o, r)),
              this.textureManager.releaseTexture(n, o, a, i)))
        var l = this.texData.get(t)
        ;(l.texture = null),
          (l.texShape = null),
          (l.isPacked = !1),
          (l.slice = null)
      }),
      (o.prototype.getTexture = function (t) {
        return this.uploadToGPU(t), this.texData.get(t).texture
      }),
      (o.prototype.getDataInfo = function (t) {
        return this.texData.get(t)
      }),
      (o.prototype.getCPUBackend = function () {
        return i().getBool('WEBGL_CPU_FORWARD')
          ? (null == this.cpuBackend &&
              (this.cpuBackend = Lt.findBackend('cpu')),
            this.cpuBackend)
          : null
      }),
      (o.prototype.shouldExecuteOnCPU = function (t, e) {
        var n = this
        return (
          void 0 === e && (e = 128),
          null != this.getCPUBackend() &&
            t.every(function (t) {
              return null == n.texData.get(t.dataId).texture && t.size < e
            })
        )
      }),
      (o.prototype.getGPGPUContext = function () {
        return this.gpgpu
      }),
      (o.prototype.complex = function (t, e) {
        var n = this.makeOutput(t.shape, 'complex64')
        return (
          (this.texData.get(n.dataId).complexTensors = {
            real: Lt.keep(t.clone()),
            imag: Lt.keep(e.clone()),
          }),
          n
        )
      }),
      (o.prototype.real = function (t) {
        return this.texData.get(t.dataId).complexTensors.real.clone()
      }),
      (o.prototype.imag = function (t) {
        return this.texData.get(t.dataId).complexTensors.imag.clone()
      }),
      (o.prototype.slice = function (t, e, n) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.slice(t, e, n)
        if (0 === k(n)) return Fn([], n, t.dtype)
        var r = this.texData.get(t.dataId).isPacked,
          o = io(t.shape, e, n)
        if (r || !o) {
          var a = i().getBool('WEBGL_PACK_ARRAY_OPERATIONS')
              ? new fs(n)
              : new ls(n),
            s = a.getCustomSetupFunc(e)
          return this.compileAndRun(a, [t], null, s)
        }
        return this.uploadToGPU(t.dataId), this.shallowSlice(t, e, n)
      }),
      (o.prototype.shallowSlice = function (t, e, n) {
        var r = this.texData.get(t.dataId),
          o = this.makeOutput(n, t.dtype),
          a = this.texData.get(o.dataId)
        Object.assign(a, r), (a.shape = n), (a.dtype = t.dtype)
        var i = so(e, t.strides)
        r.slice && (i += r.slice.flatOffset),
          (a.slice = {
            flatOffset: i,
            origDataId: (r.slice && r.slice.origDataId) || t.dataId,
          })
        var s = this.dataRefCount.get(a.slice.origDataId) || 1
        return this.dataRefCount.set(a.slice.origDataId, s + 1), o
      }),
      (o.prototype.stridedSlice = function (t, e, n, r) {
        if (this.shouldExecuteOnCPU([t]))
          return this.cpuBackend.stridedSlice(t, e, n, r)
        var o = ro(e, n, r)
        if (
          o.some(function (t) {
            return 0 === t
          })
        )
          return Fn([], o)
        var a = new ds(e, r, o)
        return this.compileAndRun(a, [t])
      }),
      (o.prototype.reverse = function (t, e) {
        var n = i().getBool('WEBGL_PACK_ARRAY_OPERATIONS')
          ? new is(t.shape, e)
          : new as(t.shape, e)
        return this.compileAndRun(n, [t])
      }),
      (o.prototype.concat = function (t, e) {
        if ('complex64' === t[0].dtype) {
          var n = t.map(function (t) {
              return Tn(t)
            }),
            r = t.map(function (t) {
              return Nn(t)
            })
          return Dn(this.concat(n, e), this.concat(r, e))
        }
        if (this.shouldExecuteOnCPU(t)) return this.cpuBackend.concat(t, e)
        if (1 === t.length) return t[0]
        if (t.length > i().getNumber('WEBGL_MAX_TEXTURES_IN_SHADER')) {
          var o = Math.floor(t.length / 2),
            a = this.concat(t.slice(0, o), e),
            s = this.concat(t.slice(o), e)
          return this.concat([a, s], e)
        }
        if (i().getBool('WEBGL_PACK_ARRAY_OPERATIONS') && t[0].rank > 1) {
          var u = new Ga(
            t.map(function (t) {
              return t.shape
            }),
            e
          )
          return this.compileAndRun(u, t)
        }
        var c = Sn(
            t.map(function (t) {
              return t.shape
            }),
            e
          ),
          l = t.map(function (t) {
            return t.as2D(-1, k(t.shape.slice(e)))
          }),
          h = new za(
            l.map(function (t) {
              return t.shape
            })
          )
        return this.compileAndRun(h, l).reshape(c)
      }),
      (o.prototype.neg = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.neg(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, Ts, t.dtype)
        var e = new Cs(t.shape, Ts)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.batchMatMul = function (t, e, n, r) {
        var o = n ? t.shape[2] : t.shape[1],
          a = r ? e.shape[1] : e.shape[2],
          i = n ? t.shape[1] : t.shape[2],
          s = t.shape[0]
        if ((1 === o || 1 === a) && i > 1e3) {
          n && (t = t.transpose([0, 2, 1])), r && (e = e.transpose([0, 2, 1]))
          var u = 1 === a ? t : t.as3D(s, i, 1),
            c = 1 === a ? 2 : 1,
            l = 1 === a ? e.as3D(s, 1, i) : e
          return this.multiply(u, l).sum(c, !0)
        }
        var h = Dt(t.dtype, e.dtype),
          f = new Hi(t.shape, [s, o, a], n, r)
        return this.compileAndRun(f, [t, e], h)
      }),
      (o.prototype.fusedBatchMatMul = function (t) {
        var e = t.a,
          n = t.b,
          r = t.transposeA,
          o = t.transposeB,
          a = t.bias,
          i = t.activation,
          s = t.preluActivationWeights,
          u = r ? e.shape[2] : e.shape[1],
          c = o ? n.shape[1] : n.shape[2],
          l = e.shape[0],
          h = Dt(e.dtype, n.dtype),
          f = null != a,
          d = null != s,
          p = i ? Qs(i, !0) : null,
          v = new Hi(e.shape, [l, u, c], r, o, f, p, d),
          m = [e, n]
        return a && m.push(a), s && m.push(s), this.compileAndRun(v, m, h)
      }),
      (o.prototype.multiply = function (t, e) {
        if ('complex64' === t.dtype) {
          var n = this.texData.get(t.dataId),
            r = this.texData.get(e.dataId),
            o = new Na(Da, t.shape, e.shape),
            a = new Na(Ta, t.shape, e.shape),
            s = [
              this.makeComplexComponentTensorInfo(t, n.complexTensors.real),
              this.makeComplexComponentTensorInfo(t, n.complexTensors.imag),
              this.makeComplexComponentTensorInfo(e, r.complexTensors.real),
              this.makeComplexComponentTensorInfo(e, r.complexTensors.imag),
            ],
            u = this.compileAndRun(o, s),
            c = this.compileAndRun(a, s),
            l = this.complex(u, c)
          return u.dispose(), c.dispose(), l
        }
        if (this.shouldExecuteOnCPU([t, e]))
          return this.cpuBackend.multiply(t, e)
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(t, e, Oa, t.dtype)
        var h = new Ba(Oa, t.shape, e.shape)
        return this.compileAndRun(h, [t, e], t.dtype)
      }),
      (o.prototype.batchNormalization = function (t, e, n, r, o, a) {
        var s = [t, e, n],
          u = null
        null != a && ((u = a.shape), s.push(a))
        var c = null
        if (
          (null != o && ((c = o.shape), s.push(o)),
          i().getBool('WEBGL_PACK_NORMALIZATION'))
        ) {
          var l = new Aa(t.shape, e.shape, n.shape, u, c, r)
          return this.compileAndRun(l, s)
        }
        var h = new Sa(t.shape, e.shape, n.shape, u, c, r)
        return this.compileAndRun(h, s)
      }),
      (o.prototype.localResponseNormalization4D = function (t, e, n, r, o) {
        var a = i().getBool('WEBGL_PACK_NORMALIZATION')
          ? new Vi(t.shape, e, n, r, o)
          : new Wi(t.shape, e, n, r, o)
        return this.compileAndRun(a, [t])
      }),
      (o.prototype.LRNGrad = function (t, e, n, r, o, a, i) {
        var s = new Ui(e.shape, r, o, a, i)
        return this.compileAndRun(s, [e, n, t])
      }),
      (o.prototype.tile = function (t, e) {
        if ('string' === t.dtype) {
          var n = this.readSync(t.dataId).map(function (t) {
            return ot(t)
          })
          return ta(dr(t.shape, t.dtype, n), e)
        }
        var r = new gs(t.shape, e)
        return this.compileAndRun(r, [t])
      }),
      (o.prototype.pad = function (t, e, n) {
        var r = i().getBool('WEBGL_PACK_ARRAY_OPERATIONS')
          ? new Yi(t.shape, e, n)
          : new Xi(t.shape, e, n)
        return this.compileAndRun(r, [t])
      }),
      (o.prototype.transpose = function (t, e) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.transpose(t, e)
        var n = i().getBool('WEBGL_PACK_ARRAY_OPERATIONS')
          ? new xs(t.shape, e)
          : new ys(t.shape, e)
        return this.compileAndRun(n, [t])
      }),
      (o.prototype.gather = function (t, e, n) {
        if (this.shouldExecuteOnCPU([t, e]))
          return this.cpuBackend.gather(t, e, n)
        var r = new mi(t.shape, e.size, n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.batchToSpaceND = function (t, e, n) {
        C(t.rank <= 4, function () {
          return 'batchToSpaceND for rank > 4 with a WebGL backend not implemented yet'
        })
        var r = e.reduce(function (t, e) {
            return t * e
          }),
          o = zr(t.shape, e, r),
          a = Gr(o.length, e.length),
          i = Hr(t.shape, e, r),
          s = qr(n, e.length),
          u = Kr(i, n, e.length)
        return t.reshape(o).transpose(a).reshape(i).slice(s, u)
      }),
      (o.prototype.spaceToBatchND = function (t, e, n) {
        C(t.rank <= 4, function () {
          return 'spaceToBatchND for rank > 4 with a WebGL backend not implemented yet'
        })
        var r = e.reduce(function (t, e) {
            return t * e
          }),
          o = [[0, 0]]
        o.push.apply(o, n)
        for (var a = 1 + e.length; a < t.shape.length; ++a) o.push([0, 0])
        var i = t.pad(o),
          s = zr(i.shape, e, r, !1),
          u = Gr(s.length, e.length, !1),
          c = Hr(i.shape, e, r, !1)
        return i.reshape(s).transpose(u).reshape(c)
      }),
      (o.prototype.reduce = function (t, e, n) {
        var r = t.shape[0],
          o = t.shape[1],
          a = $r(o),
          i = new Ji({ windowSize: a, inSize: o, batchSize: r }, e),
          s = this.compileAndRun(i, [t], n)
        return 1 === s.shape[1] ? s : this.reduce(s, e, n)
      }),
      (o.prototype.argReduce = function (t, e, n) {
        void 0 === n && (n = null)
        var r = t.shape[0],
          o = t.shape[1]
        null != n && ((r = n.shape[0]), (o = n.shape[1]))
        var a = $r(o),
          i = new aa({ windowSize: a, inSize: o, batchSize: r }, e, null == n),
          s = [t]
        null != n && s.push(n)
        var u = this.compileAndRun(i, s, 'int32')
        return 1 === u.shape[1] ? u : this.argReduce(t, e, u)
      }),
      (o.prototype.argReducePacked = function (t, e, n) {
        void 0 === n && (n = null)
        var r = null != n ? n.shape : t.shape,
          o = $r(r[r.length - 1]),
          a = new Ra(r, o, e, null == n),
          i = null == n ? [t] : [t, n],
          s = this.compileAndRun(a, i, 'int32')
        return s.rank === t.rank ? this.argReducePacked(t, e, s) : s
      }),
      (o.prototype.sum = function (t, e) {
        Cn('sum', e, t.rank)
        var n = bn(t.shape, e),
          r = n[0],
          o = k(n[1]),
          a = t.as2D(-1, o),
          i = Tt(t.dtype)
        return this.reduce(a, 'sum', i).reshape(r)
      }),
      (o.prototype.prod = function (t, e) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.prod(t, e)
        var n = bn(t.shape, e),
          r = n[0],
          o = k(n[1]),
          a = t.as2D(-1, o),
          i = Tt(t.dtype)
        return this.reduce(a, 'prod', i).reshape(r)
      }),
      (o.prototype.unsortedSegmentSum = function (t, e, n) {
        var r = 0,
          o = En([r], t.rank),
          a = t
        null != o && ((a = t.transpose(o)), (r = In(1, t.rank)[0]))
        var i = (function (t, e, n) {
            for (var r = [], o = t.length, a = 0; a < o; a++)
              a !== e ? r.push(t[a]) : r.push(n)
            return r
          })(a.shape, r, n),
          s = k([a.shape[r]]),
          u = a.as2D(-1, s),
          c = Tt(t.dtype),
          l = this.segOpCompute(u, 'unsortedSegmentSum', e, c, n).reshape(i)
        return null != o && (l = l.transpose(Rn(o))), l
      }),
      (o.prototype.segOpCompute = function (t, e, n, r, o) {
        var a = t.shape[0],
          i = t.shape[1],
          s = (function (t, e) {
            var n,
              r = !1
            for (
              t <= Yr
                ? ((n = t), (r = !0))
                : (n = Y(t, Math.floor(Math.sqrt(t))));
              !r;

            )
              n > e || n === t ? (r = !0) : (n = Y(t, n + 1))
            return n
          })(i, o),
          u = new us(
            { windowSize: s, inSize: i, batchSize: a, numSegments: o },
            e
          ),
          c = this.compileAndRun(u, [t, n], r)
        return c.shape[1] === o
          ? c
          : ((n = Kn(0, o).tile([i / s])), this.segOpCompute(c, e, n, r, o))
      }),
      (o.prototype.argMinMaxReduce = function (t, e, n) {
        var r = [e]
        if (
          (Cn('arg' + n.charAt(0).toUpperCase() + n.slice(1), r, t.rank),
          !i().getBool('WEBGL_PACK_REDUCE') || t.rank <= 2)
        ) {
          var o = bn(t.shape, r),
            a = o[0],
            s = k(o[1]),
            u = t.as2D(-1, s)
          return this.argReduce(u, n).reshape(a)
        }
        return this.argReducePacked(t, n)
      }),
      (o.prototype.argMin = function (t, e) {
        return this.argMinMaxReduce(t, e, 'min')
      }),
      (o.prototype.argMax = function (t, e) {
        return this.argMinMaxReduce(t, e, 'max')
      }),
      (o.prototype.cumsum = function (t, e, n, r) {
        if (e !== t.rank - 1)
          throw new Error(
            'WebGL cumsum shader expects an inner-most axis=' +
              (t.rank - 1) +
              ' but got axis=' +
              e
          )
        var o = new ni(t.shape, n, r)
        return this.compileAndRun(o, [t])
      }),
      (o.prototype.equal = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(equal(a, b));\n',
            'bool'
          )
        var n = new Ba('return float(a == b);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.notEqual = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(notEqual(a, b));\n',
            'bool'
          )
        var n = new Ba('return float(a != b);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.less = function (t, e) {
        if (this.shouldExecuteOnCPU([t, e])) return this.cpuBackend.less(t, e)
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(lessThan(a, b));\n',
            'bool'
          )
        var n = new Ba('return float(a < b);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.lessEqual = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(lessThanEqual(a, b));\n',
            'bool'
          )
        var n = new Ba('return float(a <= b);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.greater = function (t, e) {
        if (this.shouldExecuteOnCPU([t, e]))
          return this.cpuBackend.greater(t, e)
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(greaterThan(a, b));\n',
            'bool'
          )
        var n = new Ba('return float(a > b);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.greaterEqual = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(greaterThanEqual(a, b));\n',
            'bool'
          )
        var n = new Ba('return float(a >= b);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.logicalNot = function (t) {
        var e = new Cs(t.shape, 'return float(!(x >= 1.0));')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.logicalAnd = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return vec4(\n    vec4(greaterThanEqual(a, vec4(1.0))) *\n    vec4(greaterThanEqual(b, vec4(1.0))));\n',
            'bool'
          )
        var n = new Ba('return float(a >= 1.0 && b >= 1.0);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.logicalOr = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  return min(\n    vec4(greaterThanEqual(a, vec4(1.0))) +\n    vec4(greaterThanEqual(b, vec4(1.0))),\n    vec4(1.0));\n',
            'bool'
          )
        var n = new Ba('return float(a >= 1.0 || b >= 1.0);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e], 'bool')
      }),
      (o.prototype.select = function (t, e, n) {
        var r = new cs(t.rank, e.shape, e.rank)
        return this.compileAndRun(r, [t, e, n], Dt(e.dtype, n.dtype))
      }),
      (o.prototype.where = function (t) {
        dn(
          'tf.where() in webgl locks the UI thread. Call tf.whereAsync() instead'
        )
        var e = t.dataSync()
        return na(t.shape, e)
      }),
      (o.prototype.topk = function (t, e, n) {
        return ea(t.dataSync(), t.shape, t.dtype, e)
      }),
      (o.prototype.min = function (t, e) {
        Cn('min', e, t.rank)
        var n = bn(t.shape, e),
          r = n[0],
          o = k(n[1]),
          a = t.as2D(-1, o)
        return this.reduce(a, 'min', a.dtype).reshape(r)
      }),
      (o.prototype.minimum = function (t, e) {
        if (this.shouldExecuteOnCPU([t, e]))
          return this.cpuBackend.minimum(t, e)
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
          ? new La(
              '\n  vec4 result = vec4(min(a, b));\n  vec4 isNaN = min(vec4(isnan(a)) + vec4(isnan(b)), vec4(1.0));\n  \n  result.r = isNaN.r > 0. ? NAN : result.r;\n  result.g = isNaN.g > 0. ? NAN : result.g;\n  result.b = isNaN.b > 0. ? NAN : result.b;\n  result.a = isNaN.a > 0. ? NAN : result.a;\n\n  return result;\n',
              t.shape,
              e.shape
            )
          : new Ba(
              '\n  if (isnan(a)) return a;\n  if (isnan(b)) return b;\n\n  return min(a, b);\n',
              t.shape,
              e.shape
            )
        return this.compileAndRun(n, [t, e])
      }),
      (o.prototype.mod = function (t, e) {
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
          ? new La(
              '\n  vec4 result = mod(a, b);\n  vec4 isNaN = vec4(equal(b, vec4(0.0)));\n  \n  result.r = isNaN.r > 0. ? NAN : result.r;\n  result.g = isNaN.g > 0. ? NAN : result.g;\n  result.b = isNaN.b > 0. ? NAN : result.b;\n  result.a = isNaN.a > 0. ? NAN : result.a;\n\n  return result;\n',
              t.shape,
              e.shape
            )
          : new Ba(
              'if (b == 0.0) return NAN;\n  return mod(a, b);',
              t.shape,
              e.shape
            )
        return this.compileAndRun(n, [t, e])
      }),
      (o.prototype.max = function (t, e) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.max(t, e)
        Cn('max', e, t.rank)
        var n = bn(t.shape, e),
          r = n[0],
          o = k(n[1]),
          a = t.as2D(-1, o)
        return this.reduce(a, 'max', a.dtype).reshape(r)
      }),
      (o.prototype.maximum = function (t, e) {
        if (this.shouldExecuteOnCPU([t, e]))
          return this.cpuBackend.maximum(t, e)
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
          ? new La(
              '\n  vec4 result = vec4(max(a, b));\n  vec4 isNaN = min(vec4(isnan(a)) + vec4(isnan(b)), vec4(1.0));\n  \n  result.r = isNaN.r > 0. ? NAN : result.r;\n  result.g = isNaN.g > 0. ? NAN : result.g;\n  result.b = isNaN.b > 0. ? NAN : result.b;\n  result.a = isNaN.a > 0. ? NAN : result.a;\n\n  return result;\n',
              t.shape,
              e.shape
            )
          : new Ba(
              '\n  if (isnan(a)) return a;\n  if (isnan(b)) return b;\n\n  return max(a, b);\n',
              t.shape,
              e.shape
            )
        return this.compileAndRun(n, [t, e])
      }),
      (o.prototype.all = function (t, e) {
        Cn('all', e, t.rank)
        var n = bn(t.shape, e),
          r = n[0],
          o = k(n[1]),
          a = t.as2D(-1, o)
        return this.reduce(a, 'all', a.dtype).reshape(r)
      }),
      (o.prototype.any = function (t, e) {
        Cn('any', e, t.rank)
        var n = bn(t.shape, e),
          r = n[0],
          o = k(n[1]),
          a = t.as2D(-1, o)
        return this.reduce(a, 'any', a.dtype).reshape(r)
      }),
      (o.prototype.realDivide = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS')) {
          return this.packedBinaryOp(
            t,
            e,
            '\n  // vec4 one = vec4(equal(a, b));\n  // return one + (vec4(1.0) - one) * a / b;\n  vec4 result = a / b;\n  if(a.x == b.x) {\n    result.x = 1.;\n  }\n  if(a.y == b.y) {\n    result.y = 1.;\n  }\n  if(a.z == b.z) {\n    result.z = 1.;\n  }\n  if(a.w == b.w) {\n    result.w = 1.;\n  }\n\n  return result;\n',
            'float32',
            !0
          )
        }
        var n = new Ba(
          '\nif (a == b) {\n  return 1.0;\n};\nreturn a / b;',
          t.shape,
          e.shape
        )
        return this.compileAndRun(n, [t, e], 'float32')
      }),
      (o.prototype.floorDiv = function (t, e) {
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(
            t,
            e,
            '\n  ivec4 ia = round(a);\n  ivec4 ib = round(b);\n  bvec4 cond = notEqual(ib, ivec4(0));\n  ivec4 result = ivec4(0);\n  vec4 s = sign(a) * sign(b);\n\n  // Windows (D3D) wants guaranteed non-zero int division at compile-time.\n  if (cond[0]) {\n    result[0] = idiv(ia[0], ib[0], s[0]);\n  }\n  if (cond[1]) {\n    result[1] = idiv(ia[1], ib[1], s[1]);\n  }\n  if (cond[2]) {\n    result[2] = idiv(ia[2], ib[2], s[2]);\n  }\n  if (cond[3]) {\n    result[3] = idiv(ia[3], ib[3], s[3]);\n  }\n  return vec4(result);\n',
            'int32'
          )
        var n = new Ba(
          '\n  float s = sign(a) * sign(b);\n  int ia = round(a);\n  int ib = round(b);\n  if (ib != 0) {\n    // Windows (D3D) wants guaranteed non-zero int division at compile-time.\n    return float(idiv(ia, ib, s));\n  } else {\n    return NAN;\n  }\n',
          t.shape,
          e.shape
        )
        return this.compileAndRun(n, [t, e], 'int32')
      }),
      (o.prototype.add = function (t, e) {
        if ('complex64' === t.dtype && 'complex64' === e.dtype)
          return this.complexSeparableBinaryOp(t, e, Fa)
        if (this.shouldExecuteOnCPU([t, e])) return this.cpuBackend.add(t, e)
        var n = Dt(t.dtype, e.dtype)
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(t, e, Fa, n)
        var r = new Ba(Fa, t.shape, e.shape)
        return this.compileAndRun(r, [t, e], n)
      }),
      (o.prototype.packedUnaryOp = function (t, e, n) {
        var r = new Xs(t.shape, e)
        return this.compileAndRun(r, [t], n)
      }),
      (o.prototype.packedBinaryOp = function (t, e, n, r, o) {
        void 0 === o && (o = !1)
        var a = new La(n, t.shape, e.shape, o)
        return this.compileAndRun(a, [t, e], r)
      }),
      (o.prototype.complexSeparableBinaryOp = function (t, e, n) {
        var r = this,
          o = this.texData.get(t.dataId),
          a = this.texData.get(e.dataId),
          i = [
            [o.complexTensors.real, a.complexTensors.real],
            [o.complexTensors.imag, a.complexTensors.imag],
          ].map(function (o) {
            var a = o[0],
              i = o[1],
              s = r.makeComplexComponentTensorInfo(t, a),
              u = r.makeComplexComponentTensorInfo(e, i),
              c = new Ba(n, t.shape, e.shape)
            return r.compileAndRun(c, [s, u], Dt(a.dtype, i.dtype))
          }),
          s = i[0],
          u = i[1],
          c = this.complex(s, u)
        return s.dispose(), u.dispose(), c
      }),
      (o.prototype.makeComplexComponentTensorInfo = function (t, e) {
        return { dataId: e.dataId, dtype: e.dtype, shape: t.shape }
      }),
      (o.prototype.addN = function (t) {
        if (1 === t.length) return t[0]
        if (t.length > i().get('WEBGL_MAX_TEXTURES_IN_SHADER')) {
          var e = Math.floor(t.length / 2),
            n = this.addN(t.slice(0, e)),
            r = this.addN(t.slice(e))
          return this.addN([n, r])
        }
        var o = t
            .map(function (t) {
              return t.dtype
            })
            .reduce(function (t, e) {
              return Dt(t, e)
            }),
          a = t.map(function (t) {
            return t.shape
          }),
          s = i().getBool('WEBGL_PACK')
            ? new oa(t[0].shape, a)
            : new ra(t[0].shape, a)
        return this.compileAndRun(s, t, o)
      }),
      (o.prototype.subtract = function (t, e) {
        if ('complex64' === t.dtype && 'complex64' === e.dtype)
          return this.complexSeparableBinaryOp(t, e, _a)
        if (this.shouldExecuteOnCPU([t, e]))
          return this.cpuBackend.subtract(t, e)
        var n = Dt(t.dtype, e.dtype)
        if (i().getBool('WEBGL_PACK_BINARY_OPERATIONS'))
          return this.packedBinaryOp(t, e, _a, t.dtype)
        var r = new Ba(_a, t.shape, e.shape)
        return this.compileAndRun(r, [t, e], n)
      }),
      (o.prototype.pow = function (t, e) {
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
            ? new La(
                '\n  // isModRound1 has 1 for components with round(mod(b, 2.0)) == 1, 0 otherwise.\n  vec4 isModRound1 = vec4(equal(round(mod(b, 2.0)), ivec4(1)));\n  vec4 multiplier = sign(a) * isModRound1 + (vec4(1.0) - isModRound1);\n  vec4 result = multiplier * pow(abs(a), b);\n\n  // Ensure that a^0 = 1, including 0^0 = 1 as this correspond to TF and JS\n  bvec4 isExpZero = equal(b, vec4(0.0));\n  result.r = isExpZero.r ? 1.0 : result.r;\n  result.g = isExpZero.g ? 1.0 : result.g;\n  result.b = isExpZero.b ? 1.0 : result.b;\n  result.a = isExpZero.a ? 1.0 : result.a;\n\n  vec4 isNaN = vec4(lessThan(a, vec4(0.0))) * vec4(lessThan(floor(b), b));\n  \n  result.r = isNaN.r > 0. ? NAN : result.r;\n  result.g = isNaN.g > 0. ? NAN : result.g;\n  result.b = isNaN.b > 0. ? NAN : result.b;\n  result.a = isNaN.a > 0. ? NAN : result.a;\n\n  return result;\n',
                t.shape,
                e.shape
              )
            : new Ba(
                '\nif(a < 0.0 && floor(b) < b){\n  return NAN;\n}\nif (b == 0.0) {\n  return 1.0;\n}\nreturn (round(mod(b, 2.0)) != 1) ?\n    pow(abs(a), b) : sign(a) * pow(abs(a), b);\n',
                t.shape,
                e.shape
              ),
          r = Dt(t.dtype, e.dtype)
        return this.compileAndRun(n, [t, e], r)
      }),
      (o.prototype.ceil = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.ceil(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, Ns, t.dtype)
        var e = new Cs(t.shape, Ns)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.floor = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.floor(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, Fs, t.dtype)
        var e = new Cs(t.shape, Fs)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.sign = function (t) {
        var e = new Cs(
          t.shape,
          '\n  if (isnan(x)) { return 0.0; }\n  return sign(x);\n'
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.isNaN = function (t) {
        var e = new Cs(t.shape, 'return float(isnan(x));')
        return this.compileAndRun(e, [t], 'bool')
      }),
      (o.prototype.isInf = function (t) {
        var e = new Cs(t.shape, 'return float(isinf(x));')
        return this.compileAndRun(e, [t], 'bool')
      }),
      (o.prototype.isFinite = function (t) {
        var e = new Cs(t.shape, 'return float(!isnan(x) && !isinf(x));')
        return this.compileAndRun(e, [t], 'bool')
      }),
      (o.prototype.round = function (t) {
        var e = new Cs(
          t.shape,
          "\n  // OpenGL ES does not support round function.\n  // The algorithm is based on banker's rounding.\n  float base = floor(x);\n  if ((x - base) < 0.5) {\n    return floor(x);\n  } else if ((x - base) > 0.5) {\n    return ceil(x);\n  } else {\n    if (mod(base, 2.0) == 0.0) {\n      return base;\n    } else {\n      return base + 1.0;\n    }\n  }\n"
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.exp = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.exp(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, _s, t.dtype)
        var e = new Cs(t.shape, _s)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.expm1 = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.expm1(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, Os, t.dtype)
        var e = new Cs(t.shape, Os)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.softmax = function (t, e) {
        var n = O([e], t.shape),
          r = this.max(t, n),
          o = wn(r.shape, n),
          a = this.subtract(t, r.reshape(o)),
          i = this.exp(a),
          s = this.sum(i, n).reshape(o)
        return this.realDivide(i, s)
      }),
      (o.prototype.log = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.log(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(
            t,
            '\n  vec4 result = log(x);\n  vec4 isNaN = vec4(lessThan(x, vec4(0.0)));\n  result.r = isNaN.r == 1.0 ? NAN : result.r;\n  result.g = isNaN.g == 1.0 ? NAN : result.g;\n  result.b = isNaN.b == 1.0 ? NAN : result.b;\n  result.a = isNaN.a == 1.0 ? NAN : result.a;\n\n  return result;\n',
            t.dtype
          )
        var e = new Cs(t.shape, 'if (x < 0.0) return NAN;\n  return log(x);')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.log1p = function (t) {
        var e = new Cs(t.shape, 'return log(1.0 + x);')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.sqrt = function (t) {
        var e = new Cs(t.shape, 'return sqrt(x);')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.rsqrt = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.rsqrt(t)
        var e = new Cs(t.shape, 'return inversesqrt(x);')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.reciprocal = function (t) {
        var e = new Cs(t.shape, 'return 1.0 / x;')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.relu = function (t) {
        var e
        return (
          (e = i().getBool('WEBGL_PACK')
            ? new Xs(t.shape, qs)
            : new Cs(t.shape, ks)),
          this.compileAndRun(e, [t])
        )
      }),
      (o.prototype.relu6 = function (t) {
        var e
        return (
          (e = i().getBool('WEBGL_PACK')
            ? new Xs(t.shape, Ks)
            : new Cs(t.shape, Ss)),
          this.compileAndRun(e, [t])
        )
      }),
      (o.prototype.prelu = function (t, e) {
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
          ? new La(Pa, t.shape, e.shape)
          : new Ba(Ma, t.shape, e.shape)
        return this.compileAndRun(n, [t, e])
      }),
      (o.prototype.elu = function (t) {
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, js, t.dtype)
        var e = new Cs(t.shape, As)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.eluDer = function (t, e) {
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
          ? new La(
              '\n  vec4 bGTEZero = vec4(greaterThanEqual(b, vec4(0.)));\n  return (bGTEZero * a) + ((vec4(1.0) - bGTEZero) * (a * (b + vec4(1.0))));\n',
              t.shape,
              e.shape
            )
          : new Ba('return (b >= 1.0) ? a : a * (b + 1.0);', t.shape, e.shape)
        return this.compileAndRun(n, [t, e])
      }),
      (o.prototype.selu = function (t) {
        var e = new Cs(t.shape, Ds)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.int = function (t) {
        var e = new Cs(t.shape, 'return float(int(x));')
        return this.compileAndRun(e, [t], 'int32')
      }),
      (o.prototype.clip = function (t, e, n) {
        var r,
          o = (r = i().getBool('WEBGL_PACK_CLIP')
            ? new Ua(t.shape)
            : new Wa(t.shape)).getCustomSetupFunc(e, n)
        return this.compileAndRun(r, [t], null, o)
      }),
      (o.prototype.abs = function (t) {
        if (this.shouldExecuteOnCPU([t])) return this.cpuBackend.abs(t)
        if (i().getBool('WEBGL_PACK_UNARY_OPERATIONS'))
          return this.packedUnaryOp(t, Is, t.dtype)
        var e = new Cs(t.shape, Is)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.complexAbs = function (t) {
        var e = this.texData.get(t.dataId),
          n = new Va(t.shape),
          r = [
            this.makeComplexComponentTensorInfo(t, e.complexTensors.real),
            this.makeComplexComponentTensorInfo(t, e.complexTensors.imag),
          ]
        return this.compileAndRun(n, r)
      }),
      (o.prototype.sigmoid = function (t) {
        var e = new Cs(t.shape, 'return 1.0 / (1.0 + exp(-1.0 * x));')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.softplus = function (t) {
        var e = new Cs(
          t.shape,
          '\n  float epsilon = 1.1920928955078125e-7;\n  float threshold = log(epsilon) + 2.0;\n\n  bool too_large = x > -threshold;\n  bool too_small = x < threshold;\n\n  float result;\n  float exp_x = exp(x);\n\n  if (too_large){\n    result = x;\n  }\n  else if (too_small){\n    result = exp_x;\n  }\n  else{\n    result = log(exp_x + 1.0);\n  }\n  return result;\n'
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.sin = function (t) {
        var e = new Cs(t.shape, Ms)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.cos = function (t) {
        var e = new Cs(t.shape, Bs)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.tan = function (t) {
        var e = new Cs(t.shape, 'return tan(x);')
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.asin = function (t) {
        var e = new Cs(t.shape, Ps)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.acos = function (t) {
        var e = new Cs(t.shape, Ls)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.atan = function (t) {
        var e = new Cs(t.shape, Ws)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.atan2 = function (t, e) {
        var n = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
          ? new La(
              '\n  vec4 result = atan(a, b);\n  vec4 isNaN = min(vec4(isnan(a)) + vec4(isnan(b)), vec4(1.0));\n  \n  result.r = isNaN.r > 0. ? NAN : result.r;\n  result.g = isNaN.g > 0. ? NAN : result.g;\n  result.b = isNaN.b > 0. ? NAN : result.b;\n  result.a = isNaN.a > 0. ? NAN : result.a;\n\n  return result;\n',
              t.shape,
              e.shape
            )
          : new Ba(
              '\n  if (isnan(a)) return a;\n  if (isnan(b)) return b;\n\n  return atan(a, b);\n',
              t.shape,
              e.shape
            )
        return this.compileAndRun(n, [t, e])
      }),
      (o.prototype.sinh = function (t) {
        var e = new Cs(
          t.shape,
          '\n  float e2x = exp(x);\n  return (e2x - 1.0 / e2x) / 2.0;\n'
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.cosh = function (t) {
        var e = new Cs(
          t.shape,
          '\n  float e2x = exp(-x);\n  return (e2x + 1.0 / e2x) / 2.0;\n'
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.tanh = function (t) {
        var e = new Cs(
          t.shape,
          '\n  float e2x = exp(-2.0 * abs(x));\n  return sign(x) * (1.0 - e2x) / (1.0 + e2x);\n'
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.asinh = function (t) {
        var e = new Cs(t.shape, Us)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.acosh = function (t) {
        var e = new Cs(t.shape, Vs)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.atanh = function (t) {
        var e = new Cs(t.shape, zs)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.erf = function (t) {
        var e = new Cs(
          t.shape,
          '\n  // Error function is calculated approximately with elementary function.\n  // See "Handbook of Mathematical Functions with Formulas,\n  // Graphs, and Mathematical Tables", Abramowitz and Stegun.\n  float p = 0.3275911;\n  float a1 = 0.254829592;\n  float a2 = -0.284496736;\n  float a3 = 1.421413741;\n  float a4 = -1.453152027;\n  float a5 = 1.061405429;\n\n  float sign = sign(x);\n  x = abs(x);\n  float t = 1.0 / (1.0 + p * x);\n  return sign * (1.0 - (((((a5*t + a4)*t) + a3)*t + a2)*t + a1)*t*exp(-x*x));\n'
        )
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.step = function (t, e) {
        var n = new Cs(
          t.shape,
          (function (t) {
            return (
              void 0 === t && (t = 0),
              Es + '\n    return x > 0.0 ? 1.0 : float(' + t + ');\n  '
            )
          })(e)
        )
        return this.compileAndRun(n, [t])
      }),
      (o.prototype.conv2dByMatMul = function (t, e, n, r, o, a) {
        var s = t.shape,
          u = this.texData.get(t.dataId),
          c = n.inChannels,
          l = s[0] * s[1] * s[2],
          h = n.outChannels,
          f = 'channelsLast' === n.dataFormat,
          d = (1 === l || 1 === h) && c > 1e3,
          p = s[2] % 2 != 0 && !!u.isPacked
        if (
          d ||
          !i().getBool('WEBGL_LAZILY_UNPACK') ||
          !i().getBool('WEBGL_PACK_BINARY_OPERATIONS') ||
          !p
        ) {
          var v = f ? s[0] * s[1] * s[2] : s[0] * s[2] * s[3],
            m = this.reshape(t, [1, v, n.inChannels]),
            g = this.reshape(e, [1, n.inChannels, n.outChannels])
          return this.reshape(
            this.fusedBatchMatMul({
              a: m,
              b: g,
              transposeA: !1,
              transposeB: !1,
              bias: r,
              activation: o,
              preluActivationWeights: a,
            }),
            n.outShape
          )
        }
        var y = f ? s[0] * s[1] * (s[2] + 1) : s[0] * s[2] * (s[3] + 1),
          x = { dataId: t.dataId, shape: [1, y, n.inChannels], dtype: t.dtype },
          b = u.shape
        ;(u.shape = u.shape.slice()),
          u.shape[u.shape.length - 2]++,
          C(_e(u.shape, x.shape), function () {
            return (
              'packed reshape ' + u.shape + ' to ' + x.shape + " isn't free"
            )
          })
        var w = this.reshape(e, [1, n.inChannels, n.outChannels]),
          E = this.fusedBatchMatMul({
            a: x,
            b: w,
            transposeA: !1,
            transposeB: !1,
            bias: r,
            activation: o,
            preluActivationWeights: a,
          }),
          R = this.texData.get(E.dataId)
        return (
          C(R.isPacked, function () {
            return 'batchMatMul result is expected to be packed'
          }),
          (u.shape = b),
          (R.shape = n.outShape),
          Lt.makeTensorFromDataId(E.dataId, n.outShape, E.dtype)
        )
      }),
      (o.prototype.conv2dWithIm2Row = function (t, e, n, r, o, a) {
        var i = n.filterWidth,
          s = n.filterHeight,
          u = n.inChannels,
          c = n.outWidth,
          l = n.outHeight,
          h = 'channelsLast' === n.dataFormat,
          f = i * s * u,
          d = l * c,
          p = [f, d],
          v = t.squeeze([0]),
          m = e.reshape([1, f, -1]),
          g = new Li(p, v.shape, n),
          y = this.compileAndRun(g, [v]).reshape([1, p[0], p[1]]),
          x = null != r,
          b = null != a,
          w = o ? Qs(o, !0) : null,
          C = new Hi(y.shape, [1, d, n.outChannels], !0, !1, x, w, b),
          E = [y, m]
        r && E.push(r), b && E.push(a)
        var R = this.compileAndRun(C, E)
        return h
          ? R.reshape([1, l, c, n.outChannels])
          : R.reshape([1, n.outChannels, l, c])
      }),
      (o.prototype.fusedConv2d = function (t) {
        var e = t.input,
          n = t.filter,
          r = t.convInfo,
          o = t.bias,
          a = t.activation,
          s = t.preluActivationWeights
        if (
          1 === r.filterHeight &&
          1 === r.filterWidth &&
          1 === r.dilationHeight &&
          1 === r.dilationWidth &&
          1 === r.strideHeight &&
          1 === r.strideWidth &&
          ('SAME' === r.padInfo.type || 'VALID' === r.padInfo.type)
        )
          return this.conv2dByMatMul(e, n, r, o, a, s)
        if (i().getBool('WEBGL_CONV_IM2COL') && 1 === e.shape[0])
          return this.conv2dWithIm2Row(e, n, r, o, a, s)
        var u = null != o,
          c = null != s,
          l = a ? Qs(a, !1) : null,
          h = new Qa(r, u, l, c),
          f = [e, n]
        return o && f.push(o), s && f.push(s), this.compileAndRun(h, f)
      }),
      (o.prototype.conv2d = function (t, e, n) {
        if (
          1 === n.filterHeight &&
          1 === n.filterWidth &&
          1 === n.dilationHeight &&
          1 === n.dilationWidth &&
          1 === n.strideHeight &&
          1 === n.strideWidth &&
          ('SAME' === n.padInfo.type || 'VALID' === n.padInfo.type)
        )
          return this.conv2dByMatMul(t, e, n)
        if (i().getBool('WEBGL_CONV_IM2COL') && 1 === t.shape[0])
          return this.conv2dWithIm2Row(t, e, n)
        var r = new Qa(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.conv2dDerInput = function (t, e, n) {
        var r = new Ka(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.conv2dDerFilter = function (t, e, n) {
        var r = new qa(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.fusedDepthwiseConv2D = function (t) {
        var e,
          n = t.input,
          r = t.filter,
          o = t.convInfo,
          a = t.bias,
          s = t.activation,
          u = t.preluActivationWeights,
          c =
            i().getBool('WEBGL_PACK_DEPTHWISECONV') &&
            o.strideWidth <= 2 &&
            o.outChannels / o.inChannels == 1,
          l = s ? Qs(s, c) : null,
          h = [n, r],
          f = null != a,
          d = null != u
        return (
          f && h.push(a),
          d && h.push(u),
          c
            ? ((e = new ti(o, f, l, d)), this.compileAndRun(e, h))
            : ((e = new Za(o, f, l, d)), this.compileAndRun(e, h))
        )
      }),
      (o.prototype.depthwiseConv2D = function (t, e, n) {
        var r
        return i().getBool('WEBGL_PACK_DEPTHWISECONV') &&
          n.strideWidth <= 2 &&
          n.outChannels / n.inChannels == 1
          ? ((r = new ti(n)), this.compileAndRun(r, [t, e]))
          : ((r = new Za(n)), this.compileAndRun(r, [t, e]))
      }),
      (o.prototype.depthwiseConv2DDerInput = function (t, e, n) {
        var r = new $a(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.depthwiseConv2DDerFilter = function (t, e, n) {
        var r = new Ya(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.conv3d = function (t, e, n) {
        var r = new Ja(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.conv3dDerInput = function (t, e, n) {
        var r = new Xa(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.conv3dDerFilter = function (t, e, n) {
        var r = new ja(n)
        return this.compileAndRun(r, [t, e])
      }),
      (o.prototype.maxPool = function (t, e) {
        var n = new $i(e, 'max', !1)
        return this.compileAndRun(n, [t])
      }),
      (o.prototype.avgPool = function (t, e) {
        var n = new $i(e, 'avg', !1)
        return this.compileAndRun(n, [t], 'float32')
      }),
      (o.prototype.maxPoolBackprop = function (t, e, n, r) {
        var o = new $i(r, 'max', !0),
          a = this.compileAndRun(o, [e]),
          i = new zi(r),
          s = this.compileAndRun(i, [t, a], e.dtype)
        return a.dispose(), s
      }),
      (o.prototype.avgPoolBackprop = function (t, e, n) {
        var r = new Ia(n)
        return this.compileAndRun(r, [t], e.dtype)
      }),
      (o.prototype.cast = function (t, e) {
        return Po(t, e, this)
      }),
      (o.prototype.unstack = function (t, e) {
        for (
          var n = t.shape[e], r = new Array(t.rank - 1), o = 0, a = 0;
          a < t.rank;
          a++
        )
          a !== e && (r[o++] = t.shape[a])
        var i = new Array(t.rank).fill(0),
          s = t.shape.slice()
        s[e] = 1
        var u = new Array(n)
        for (a = 0; a < u.length; a++)
          (i[e] = a), (u[a] = this.slice(t, i, s).reshape(r))
        return u
      }),
      (o.prototype.avgPool3d = function (t, e) {
        var n = new Qi(e, 'avg', !1)
        return this.compileAndRun(n, [t], 'float32')
      }),
      (o.prototype.avgPool3dBackprop = function (t, e, n) {
        var r = new ka(n)
        return this.compileAndRun(r, [t], e.dtype)
      }),
      (o.prototype.maxPool3d = function (t, e) {
        var n = new Qi(e, 'max', !1)
        return this.compileAndRun(n, [t], 'float32')
      }),
      (o.prototype.maxPool3dBackprop = function (t, e, n, r) {
        var o = new Qi(r, 'max', !0),
          a = this.compileAndRun(o, [e]),
          i = new Gi(r),
          s = this.compileAndRun(i, [t, a], e.dtype)
        return a.dispose(), s
      }),
      (o.prototype.reshape = function (t, e) {
        var n = this.texData.get(t.dataId)
        if (
          n.isPacked &&
          !_e(t.shape, e) &&
          (null === n.texture || !_e(n.shape, e))
        ) {
          var r = this.packedReshape(t, e)
          return Lt.makeTensorFromDataId(r.dataId, r.shape, r.dtype)
        }
        return Lo(t, e)
      }),
      (o.prototype.resizeBilinear = function (t, e, n, r) {
        var o = i().getBool('WEBGL_PACK_IMAGE_OPERATIONS')
          ? new ns(t.shape, e, n, r)
          : new es(t.shape, e, n, r)
        return this.compileAndRun(o, [t], 'float32')
      }),
      (o.prototype.resizeBilinearBackprop = function (t, e, n) {
        var r = new ts(t, e, n)
        return this.compileAndRun(r, [t])
      }),
      (o.prototype.resizeNearestNeighbor = function (t, e, n, r) {
        var o = new os(t.shape, e, n, r)
        return this.compileAndRun(o, [t])
      }),
      (o.prototype.resizeNearestNeighborBackprop = function (t, e, n) {
        var r = new rs(t, e, n)
        return this.compileAndRun(r, [t])
      }),
      (o.prototype.multinomial = function (t, e, n, r) {
        var o = e ? t : go(t),
          a = o.shape[0],
          i = o.shape[1],
          s = new qi(a, i, n),
          u = s.getCustomSetupFunc(r)
        return this.compileAndRun(s, [o], 'int32', u)
      }),
      (o.prototype.oneHot = function (t, e, n, r) {
        var o = new Ki(t.size, e, n, r)
        return this.compileAndRun(o, [t])
      }),
      (o.prototype.diag = function (t) {
        var e = new si(t.size)
        return this.compileAndRun(e, [t])
      }),
      (o.prototype.nonMaxSuppression = function (t, e, n, r, o) {
        return (
          dn(
            'tf.nonMaxSuppression() in webgl locks the UI thread. Call tf.nonMaxSuppressionAsync() instead'
          ),
          jo(t.dataSync(), e.dataSync(), n, r, o)
        )
      }),
      (o.prototype.cropAndResize = function (t, e, n, r, o, a) {
        var i = new ei(t.shape, e.shape, r, o, a)
        return this.compileAndRun(i, [t, e, n], 'float32')
      }),
      (o.prototype.depthToSpace = function (t, e, n) {
        C(e > 1, function () {
          return 'blockSize should be > 1 for depthToSpace, but was: ' + e
        })
        var r = t.shape[0],
          o = 'NHWC' === n ? t.shape[1] : t.shape[2],
          a = 'NHWC' === n ? t.shape[2] : t.shape[3],
          i = 'NHWC' === n ? t.shape[3] : t.shape[1],
          s = o * e,
          u = a * e,
          c = i / (e * e),
          l = new ii('NHWC' === n ? [r, s, u, c] : [r, c, s, u], e, n)
        return this.compileAndRun(l, [t])
      }),
      (o.prototype.split = function (t, e, n) {
        return Zo(t, e, n)
      }),
      (o.prototype.scatterND = function (t, e, n) {
        var r = Zr(0, t, n),
          o = r.sliceRank,
          a = r.numUpdates,
          i = r.sliceSize,
          s = r.strides,
          u = r.outputSize,
          c = [u / i, i],
          l = t.reshape([a, o]),
          h = e.reshape([a, i])
        if (0 === u) return Lo(Fn([]), n)
        var f = On(0),
          d = new ss(a, o, l.rank, h.rank, s, c)
        return this.compileAndRun(d, [h, l, f]).reshape(n)
      }),
      (o.prototype.sparseToDense = function (t, e, n, r) {
        var o = Zr(0, t, n),
          a = o.sliceRank,
          i = o.numUpdates,
          s = o.strides,
          u = o.outputSize,
          c = new ss(i, a, t.rank, e.rank, s, [u, 1], !1)
        return this.compileAndRun(c, [e, t, r]).reshape(n)
      }),
      (o.prototype.fft = function (t) {
        return this.fftImpl(t, !1)
      }),
      (o.prototype.ifft = function (t) {
        return this.fftImpl(t, !0)
      }),
      (o.prototype.fftImpl = function (t, e) {
        var n = this.texData.get(t.dataId),
          r = new pi(fi, t.shape, e),
          o = new pi(di, t.shape, e),
          a = [
            this.makeComplexComponentTensorInfo(t, n.complexTensors.real),
            this.makeComplexComponentTensorInfo(t, n.complexTensors.imag),
          ],
          i = this.compileAndRun(r, a),
          s = this.compileAndRun(o, a),
          u = this.complex(i, s).as2D(t.shape[0], t.shape[1])
        return i.dispose(), s.dispose(), u
      }),
      (o.prototype.gatherND = function (t, e) {
        var n = e.shape,
          r = n[n.length - 1],
          o = jr(t, e),
          a = o[0],
          i = o[1],
          s = o[2],
          u = o[3],
          c = e.reshape([i, r]),
          l = t.reshape([t.size / s, s]),
          h = new gi(r, u, [i, s])
        return this.compileAndRun(h, [l, c]).reshape(a)
      }),
      (o.prototype.fill = function (t, e, n) {
        if ('string' === (n = n || j(e))) {
          var r = P(n, k(t))
          return r.fill(e), Lt.makeTensor(r, t, n, this)
        }
        var o = new vi(t, e),
          a = o.getCustomSetupFunc(e)
        return this.compileAndRun(o, [], n, a)
      }),
      (o.prototype.onesLike = function (t) {
        if ('string' === t.dtype)
          throw new Error('onesLike is not supported under string dtype')
        return this.fill(t.shape, 1, t.dtype)
      }),
      (o.prototype.zerosLike = function (t) {
        return this.fill(t.shape, 'string' === t.dtype ? '' : 0, t.dtype)
      }),
      (o.prototype.linspace = function (t, e, n) {
        return Wo(t, e, n)
      }),
      (o.prototype.makeTensorInfo = function (t, e) {
        var n = this.write(null, t, e)
        return (
          (this.texData.get(n).usage = null), { dataId: n, shape: t, dtype: e }
        )
      }),
      (o.prototype.makeOutput = function (t, e) {
        var n = this.makeTensorInfo(t, e).dataId
        return Lt.makeTensorFromDataId(n, t, e, this)
      }),
      (o.prototype.unpackTensor = function (t) {
        var e = new Ys(t.shape)
        return this.runWebGLProgram(e, [t], t.dtype)
      }),
      (o.prototype.packTensor = function (t) {
        var e = new ji(t.shape)
        return this.runWebGLProgram(e, [t], t.dtype, null, !0)
      }),
      (o.prototype.packedReshape = function (t, e) {
        var n = [Ae(t.shape)].concat(De(t.shape)),
          r = { dtype: t.dtype, shape: n, dataId: t.dataId },
          o = [Ae(e)].concat(De(e)),
          a = new Zi(o, n),
          i = this.runWebGLProgram(a, [r], t.dtype, null, !0)
        return { dataId: i.dataId, shape: e, dtype: i.dtype }
      }),
      (o.prototype.decode = function (t) {
        var e,
          n = this.texData.get(t),
          r = n.isPacked,
          o = n.shape,
          a = n.dtype,
          i = Te(o)
        e = r ? new ai(i) : new oi(i)
        return {
          dtype: a,
          shape: o,
          dataId: this.runWebGLProgram(
            e,
            [{ shape: i, dtype: a, dataId: t }],
            a,
            null,
            !0
          ).dataId,
        }
      }),
      (o.prototype.runWebGLProgram = function (t, e, n, r, o) {
        var a = this
        void 0 === o && (o = !1)
        var s = this.makeTensorInfo(t.outputShape, n),
          u = this.texData.get(s.dataId)
        if (
          (t.packedOutput && (u.isPacked = !0), t.outPackingScheme === Vt.DENSE)
        ) {
          var c = Yt(t.outputShape)
          u.texShape = c.map(function (t) {
            return 2 * t
          })
        }
        if (
          (null != t.outTexUsage && (u.usage = t.outTexUsage), 0 === k(s.shape))
        )
          return (u.values = B(s.dtype, 0)), s
        var l = [],
          h = e.map(function (e) {
            if ('complex64' === e.dtype)
              throw new Error(
                'GPGPUProgram does not support complex64 input. For complex64 dtypes, please separate the program into real and imaginary parts.'
              )
            var n = a.texData.get(e.dataId)
            if (null == n.texture) {
              if (
                !t.packedInputs &&
                k(e.shape) <= i().getNumber('WEBGL_SIZE_UPLOAD_UNIFORM')
              )
                return {
                  shape: e.shape,
                  texData: null,
                  isUniform: !0,
                  uniformValues: n.values,
                }
              t.packedInputs && ((n.isPacked = !0), (n.shape = e.shape))
            } else if (!!n.isPacked != !!t.packedInputs)
              (e = n.isPacked ? a.unpackTensor(e) : a.packTensor(e)),
                l.push(e),
                (n = a.texData.get(e.dataId))
            else if (n.isPacked && !_e(n.shape, e.shape)) {
              var r = e,
                o = e.shape
              ;(e.shape = n.shape),
                (e = a.packedReshape(e, o)),
                l.push(e),
                (n = a.texData.get(e.dataId)),
                (r.shape = o)
            }
            return (
              a.uploadToGPU(e.dataId),
              { shape: e.shape, texData: n, isUniform: !1 }
            )
          })
        this.uploadToGPU(s.dataId)
        var f,
          d = { shape: s.shape, texData: u, isUniform: !1 },
          p = (function (t, e, n) {
            var r = ''
            e.concat(n).forEach(function (t) {
              var e =
                  null != t.texData &&
                  null != t.texData.slice &&
                  t.texData.slice.flatOffset > 0,
                n = t.isUniform ? 'uniform' : t.texData.texShape
              r += t.shape + '_' + n + '_' + e
            })
            var o = t.userCode,
              a = t.constructor.name
            return (a += '_' + r + '_' + o)
          })(t, h, d),
          v = this.getAndSaveBinary(p, function () {
            return (function (t, e, n, r) {
              var o = e.userCode,
                a = n.map(function (t, n) {
                  var r = {
                    logicalShape: t.shape,
                    texShape: t.isUniform ? null : t.texData.texShape,
                    isUniform: t.isUniform,
                    isPacked: !t.isUniform && t.texData.isPacked,
                    flatOffset: null,
                  }
                  return (
                    null != t.texData &&
                      null != t.texData.slice &&
                      t.texData.slice.flatOffset > 0 &&
                      (r.flatOffset = t.texData.slice.flatOffset),
                    { name: e.variableNames[n], shapeInfo: r }
                  )
                }),
                s = a.map(function (t) {
                  return t.shapeInfo
                }),
                u = {
                  logicalShape: r.shape,
                  texShape: r.texData.texShape,
                  isUniform: !1,
                  isPacked: r.texData.isPacked,
                  flatOffset: null,
                },
                c = fa(a, u, o, e.packedInputs),
                l = t.createProgram(c),
                h = null,
                f = t.getUniformLocation(l, 'NAN', !1)
              1 === i().getNumber('WEBGL_VERSION') &&
                (h = t.getUniformLocation(l, 'INFINITY', !1))
              for (var d = {}, p = 0; p < e.variableNames.length; p++) {
                var v = e.variableNames[p]
                ;(d[v] = t.getUniformLocation(l, v, !1)),
                  (d['offset' + v] = t.getUniformLocation(l, 'offset' + v, !1))
              }
              return {
                program: e,
                source: c,
                webGLProgram: l,
                uniformLocations: d,
                inShapeInfos: s,
                outShapeInfo: u,
                infLoc: h,
                nanLoc: f,
              }
            })(a.gpgpu, t, h, d)
          }),
          m = null != this.activeTimers
        if (
          (m && (f = this.startTimer()),
          (function (t, e, n, r, o) {
            Pi(e.inShapeInfos, n), Pi([e.outShapeInfo], [r])
            var a = r.texData.texture,
              s = r.texData.texShape
            r.texData.isPacked
              ? t.setOutputPackedMatrixTexture(a, s[0], s[1])
              : t.setOutputMatrixTexture(a, s[0], s[1]),
              t.setProgram(e.webGLProgram),
              1 === i().getNumber('WEBGL_VERSION') &&
                null !== e.infLoc &&
                t.gl.uniform1f(e.infLoc, 1 / 0),
              null !== e.nanLoc && t.gl.uniform1f(e.nanLoc, NaN),
              n.forEach(function (n, r) {
                var o = e.program.variableNames[r],
                  a = e.uniformLocations[o],
                  i = e.uniformLocations['offset' + o]
                if (null != a)
                  if (n.isUniform)
                    if (k(n.shape) < 2) t.gl.uniform1f(a, n.uniformValues[0])
                    else {
                      var s = n.uniformValues
                      s instanceof Float32Array || (s = new Float32Array(s)),
                        t.gl.uniform1fv(a, s)
                    }
                  else
                    null != n.texData.slice &&
                      null != i &&
                      t.gl.uniform1i(i, n.texData.slice.flatOffset),
                      t.setInputMatrixTexture(n.texData.texture, a, r)
              }),
              null != o && o(t, e.webGLProgram),
              t.executeProgram()
          })(this.gpgpu, v, h, d, r),
          l.forEach(function (t) {
            return a.disposeData(t.dataId)
          }),
          m &&
            ((f = this.endTimer(f)),
            this.activeTimers.push({
              name: t.constructor.name,
              query: this.getQueryTime(f),
            })),
          !i().getBool('WEBGL_LAZILY_UNPACK') && u.isPacked && !1 === o)
        ) {
          var g = this.unpackTensor(s)
          return this.disposeData(s.dataId), g
        }
        return s
      }),
      (o.prototype.compileAndRun = function (t, e, n, r, o) {
        void 0 === o && (o = !1), (n = n || e[0].dtype)
        var a = this.runWebGLProgram(t, e, n, r, o)
        return Lt.makeTensorFromDataId(a.dataId, a.shape, a.dtype)
      }),
      (o.prototype.getAndSaveBinary = function (t, e) {
        return (
          t in this.binaryCache || (this.binaryCache[t] = e()),
          this.binaryCache[t]
        )
      }),
      (o.prototype.getTextureManager = function () {
        return this.textureManager
      }),
      (o.prototype.dispose = function () {
        var t = this
        if (!this.disposed) {
          if (!i().getBool('IS_TEST'))
            Object.keys(this.binaryCache).forEach(function (e) {
              t.gpgpu.deleteProgram(t.binaryCache[e].webGLProgram),
                delete t.binaryCache[e]
            })
          this.textureManager.dispose(),
            null != this.canvas &&
            'undefined' != typeof HTMLCanvasElement &&
            this.canvas instanceof HTMLCanvasElement
              ? this.canvas.remove()
              : (this.canvas = null),
            this.gpgpuCreatedLocally &&
              ((this.gpgpu.program = null), this.gpgpu.dispose()),
            (this.disposed = !0)
        }
      }),
      (o.prototype.floatPrecision = function () {
        var t = this
        return (
          null == this.floatPrecisionValue &&
            (this.floatPrecisionValue = Ze(function () {
              if (!i().get('WEBGL_RENDER_FLOAT32_ENABLED')) {
                var e = i().getBool('DEBUG')
                i().set('DEBUG', !1)
                var n = t.abs(On(1e-8)).dataSync()[0]
                if ((i().set('DEBUG', e), n > 0)) return 32
              }
              return 16
            })),
          this.floatPrecisionValue
        )
      }),
      (o.prototype.epsilon = function () {
        return 32 === this.floatPrecision() ? 1e-7 : 1e-4
      }),
      (o.prototype.uploadToGPU = function (t) {
        var e,
          n = this.texData.get(t),
          r = n.shape,
          o = n.dtype,
          a = n.values,
          i = n.texture,
          s = n.usage,
          u = n.isPacked
        if (null == i) {
          var c,
            l = null != this.activeTimers
          l && (c = et())
          var h = n.texShape
          if ((null == h && ((h = Ne(r, u)), (n.texShape = h)), null != a)) {
            var f = Te(r),
              d = void 0,
              p = h[1],
              v = h[0],
              m = a instanceof Uint8Array
            u
              ? ((p = (e = $t(h[0], h[1]))[0]),
                (v = e[1]),
                (d = new hi(f, [v, p], m)))
              : (d = new li(f, [v, p], m))
            var g = this.makeTensorInfo([v, p], o)
            ;(this.texData.get(g.dataId).usage = m ? zt.PIXELS : zt.UPLOAD),
              this.gpgpu.uploadDenseMatrixToTexture(
                this.getTexture(g.dataId),
                p,
                v,
                a
              )
            var y = this.runWebGLProgram(d, [g], o, null, !0),
              x = this.texData.get(y.dataId)
            ;(n.texture = x.texture),
              (n.texShape = x.texShape),
              (n.isPacked = x.isPacked),
              (n.usage = x.usage),
              this.disposeData(g.dataId),
              this.texData.delete(y.dataId),
              (n.values = null),
              l && (this.uploadWaitMs += et() - c)
          } else {
            var b = this.acquireTexture(h, s, o, u)
            n.texture = b
          }
        }
      }),
      (o.prototype.convertAndCacheOnCPU = function (t, e) {
        var n = this.texData.get(t),
          r = n.dtype
        return (
          this.releaseGPUData(t),
          null != e &&
            (n.values = (function (t, e) {
              if ('float32' === e || 'complex64' === e) return t
              if ('int32' === e || 'bool' === e) {
                for (
                  var n =
                      'int32' === e
                        ? new Int32Array(t.length)
                        : new Uint8Array(t.length),
                    r = 0;
                  r < n.length;
                  ++r
                )
                  n[r] = Math.round(t[r])
                return n
              }
              throw new Error('Unknown dtype ' + e)
            })(e, r)),
          n.values
        )
      }),
      (o.prototype.acquireTexture = function (t, e, n, r) {
        if (
          ((this.numBytesInGPU += this.computeBytes(t, n)),
          !this.warnedAboutMemory &&
            this.numBytesInGPU > 1024 * this.numMBBeforeWarning * 1024)
        ) {
          var o = (this.numBytesInGPU / 1024 / 1024).toFixed(2)
          ;(this.warnedAboutMemory = !0),
            console.warn(
              'High memory usage in GPU: ' +
                o +
                ' MB, most likely due to a memory leak'
            )
        }
        return this.textureManager.acquireTexture(t, e, r)
      }),
      (o.prototype.computeBytes = function (t, e) {
        return t[0] * t[1] * z(e)
      }),
      o
    )
  })(bo)
  Wt() &&
    Lt.registerBackend(
      'webgl',
      function () {
        return new Zs()
      },
      2
    )
  var tu = An({
      square_: function (t) {
        var e = mn(t, 'x', 'square'),
          n = [e]
        return Lt.runKernelFunc(
          function (t, n) {
            return n([e]), t.square(e)
          },
          { x: e },
          null,
          'Square',
          {},
          n,
          []
        )
      },
    }),
    eu = 'SquaredDifference'
  var nu = An({
    squaredDifference_: function (t, e) {
      var n,
        r = mn(t, 'a', 'squaredDifference'),
        o = mn(e, 'b', 'squaredDifference')
      ;(n = Nt(r, o)), (r = n[0]), (o = n[1]), Ro(r.shape, o.shape)
      var a = { a: r, b: o },
        i = [r, o]
      return Lt.runKernelFunc(
        function (t, e) {
          var n = t.squaredDifference(r, o)
          return e([r, o]), n
        },
        a,
        function (t, e) {
          var n = e[0],
            r = e[1],
            o = On(2)
          return {
            a: function () {
              return t.mul(n.sub(r).mul(o))
            },
            b: function () {
              return t.mul(r.sub(n).mul(o))
            },
          }
        },
        eu,
        {},
        i,
        []
      )
    },
  })
  var ru = An({
      abs_: function (t) {
        var e = mn(t, 'x', 'abs')
        return 'complex64' === e.dtype
          ? Lt.runKernelFunc(
              function (t) {
                return t.complexAbs(e)
              },
              { $x: e }
            )
          : Lt.runKernelFunc(
              function (t, n) {
                var r = t.abs(e)
                return n([e]), r
              },
              { x: e },
              function (t, e) {
                var n = e[0]
                return {
                  x: function () {
                    return t.mul(n.toFloat().step(-1))
                  },
                }
              },
              'Abs'
            )
      },
    }),
    ou = An({
      acos_: function (t) {
        var e = mn(t, 'x', 'acos')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.acos(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.divStrict(On(1).sub(n.toFloat().square()).sqrt()).neg()
              },
            }
          }
        )
      },
    }),
    au = An({
      acosh_: function (t) {
        var e = mn(t, 'x', 'acosh')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.acosh(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.divStrict(n.toFloat().square().sub(1).sqrt())
              },
            }
          }
        )
      },
    }),
    iu = An({
      asin_: function (t) {
        var e = mn(t, 'x', 'asin')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.asin(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.divStrict(On(1).sub(n.toFloat().square()).sqrt())
              },
            }
          }
        )
      },
    }),
    su = An({
      asinh_: function (t) {
        var e = mn(t, 'x', 'asinh')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.asinh(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.divStrict(On(1).add(n.toFloat().square()).sqrt())
              },
            }
          }
        )
      },
    }),
    uu = An({
      atan_: function (t) {
        var e = mn(t, 'x', 'atan')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.atan(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.div(n.toFloat().square().add(1))
              },
            }
          }
        )
      },
    }),
    cu = An({
      atanh_: function (t) {
        var e = mn(t, 'x', 'atanh')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.atanh(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.div(On(1).sub(n.toFloat().square()))
              },
            }
          }
        )
      },
    }),
    lu = An({
      ceil_: function (t) {
        var e = mn(t, 'x', 'ceil')
        return Lt.runKernelFunc(
          function (t) {
            return t.ceil(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    hu = An({
      clipByValue_: function (t, e, n) {
        var r = mn(t, 'x', 'clipByValue')
        C(e <= n, function () {
          return (
            'Error in clip: min (' +
            e +
            ') must be less than or equal to max (' +
            n +
            ').'
          )
        })
        var o = [r],
          a = { min: e, max: n }
        return Lt.runKernelFunc(
          function (t, o) {
            var a = t.clip(r, e, n)
            return o([r]), a
          },
          { x: r },
          function (t, r) {
            var o = r[0]
            return {
              x: function () {
                return t.where(
                  o.greaterEqual(e).logicalAnd(o.lessEqual(n)),
                  Xn(t)
                )
              },
            }
          },
          'ClipByValue',
          a,
          o
        )
      },
    }),
    fu = An({
      cos_: function (t) {
        var e = mn(t, 'x', 'cos'),
          n = [e]
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.cos(e)
            return n([e]), r
          },
          { x: e },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return n.toFloat().sin().neg().mul(t)
              },
            }
          },
          'Cos',
          {},
          n
        )
      },
    }),
    du = An({
      cosh_: function (t) {
        var e = mn(t, 'x', 'cosh')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.cosh(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return n.toFloat().sinh().mulStrict(t)
              },
            }
          }
        )
      },
    }),
    pu = An({
      erf_: function (t) {
        var e = mn(t, 'x', 'erf')
        return (
          C('int32' === e.dtype || 'float32' === e.dtype, function () {
            return 'Input dtype must be `int32` or `float32`.'
          }),
          'int32' === e.dtype && (e = e.toFloat()),
          Lt.runKernelFunc(
            function (t, n) {
              var r = t.erf(e)
              return n([e]), r
            },
            { $x: e },
            function (t, e) {
              var n = e[0]
              return {
                $x: function () {
                  return t.mul(
                    n
                      .square()
                      .neg()
                      .exp()
                      .mul(2 / Math.sqrt(Math.PI))
                  )
                },
              }
            }
          )
        )
      },
    }),
    vu = An({
      exp_: function (t) {
        var e = mn(t, 'x', 'exp')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.exp(e)
            return n([r]), r
          },
          { x: e },
          function (t, e) {
            return {
              x: function () {
                return t.mulStrict(e[0])
              },
            }
          },
          'Exp',
          {},
          [],
          [!0]
        )
      },
    }),
    mu = An({
      expm1_: function (t) {
        var e = mn(t, 'x', 'expm1')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.expm1(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.mul(n.exp())
              },
            }
          }
        )
      },
    }),
    gu = An({
      floor_: function (t) {
        var e = mn(t, 'x', 'floor')
        return Lt.runKernelFunc(
          function (t) {
            return t.floor(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    yu = An({
      log_: function (t) {
        var e = mn(t, 'x', 'log'),
          n = [e]
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.log(e)
            return n([e]), r
          },
          { x: e },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return t.div(n.toFloat())
              },
            }
          },
          'Log',
          {},
          n
        )
      },
    }),
    xu = An({
      log1p_: function (t) {
        var e = mn(t, 'x', 'log1p')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.log1p(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.div(n.add(1))
              },
            }
          }
        )
      },
    }),
    bu = An({
      logSigmoid_: function (t) {
        var e = mn(t, 'x', 'logSigmoid')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.softplus(e.neg()).neg()
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.mul(n.neg().sigmoid())
              },
            }
          }
        )
      },
    }),
    wu = An({
      neg_: function (t) {
        var e = mn(t, 'x', 'neg'),
          n = [e]
        return Lt.runKernelFunc(
          function (t) {
            return t.neg(e)
          },
          { x: e },
          function (t) {
            return {
              x: function () {
                return t.neg()
              },
            }
          },
          'Neg',
          {},
          n
        )
      },
    }),
    Cu = An({
      reciprocal_: function (t) {
        var e = mn(t, 'x', 'reciprocal')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.reciprocal(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.div(n.square().neg())
              },
            }
          }
        )
      },
    }),
    Eu = An({
      round_: function (t) {
        var e = mn(t, 'x', 'round')
        return Lt.runKernelFunc(
          function (t) {
            return t.round(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Ru = An({
      rsqrt_: function (t) {
        var e = mn(t, 'x', 'rsqrt'),
          n = [e]
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.rsqrt(e)
            return n([e]), r
          },
          { x: e },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return t.div(n.pow(1.5).mul(2)).neg()
              },
            }
          },
          'Rsqrt',
          {},
          n
        )
      },
    }),
    Iu = An({
      sigmoid_: function (t) {
        var e = mn(t, 'x', 'sigmoid')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.sigmoid(e)
            return n([r]), r
          },
          { x: e },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return t.mul(n.mul(On(1).sub(n)))
              },
            }
          },
          'Sigmoid'
        )
      },
    }),
    ku = An({
      sign_: function (t) {
        var e = mn(t, 'x', 'sign')
        return Lt.runKernelFunc(
          function (t) {
            return t.sign(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Su = An({
      isNaN_: function (t) {
        var e = mn(t, 'x', 'isNaN')
        return Lt.runKernelFunc(
          function (t) {
            return t.isNaN(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Au = An({
      isInf_: function (t) {
        var e = mn(t, 'x', 'isInf')
        return Lt.runKernelFunc(
          function (t) {
            return t.isInf(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Du = An({
      isFinite_: function (t) {
        var e = mn(t, 'x', 'isFinite')
        return Lt.runKernelFunc(
          function (t) {
            return t.isFinite(e)
          },
          { $x: e },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Tu = An({
      sin_: function (t) {
        var e = mn(t, 'x', 'sin'),
          n = [e]
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.sin(e)
            return n([e]), r
          },
          { x: e },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return n.toFloat().cos().mul(t)
              },
            }
          },
          'Sin',
          {},
          n
        )
      },
    }),
    Nu = An({
      sinh_: function (t) {
        var e = mn(t, 'x', 'sinh')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.sinh(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return n.toFloat().cosh().mulStrict(t)
              },
            }
          }
        )
      },
    }),
    Fu = An({
      softplus_: function (t) {
        var e = mn(t, 'x', 'softplus')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.softplus(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.mul(n.sigmoid())
              },
            }
          }
        )
      },
    }),
    _u = An({
      sqrt_: function (t) {
        var e = mn(t, 'x', 'sqrt')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.sqrt(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.div(n.toFloat().sqrt().mul(2))
              },
            }
          }
        )
      },
    }),
    Ou = An({
      step_: function (t, e) {
        void 0 === e && (e = 0)
        var n = mn(t, 'x', 'step')
        return Lt.runKernelFunc(
          function (t) {
            return t.step(n, e)
          },
          { $x: n },
          function (t) {
            return {
              $x: function () {
                return Xn(t)
              },
            }
          }
        )
      },
    }),
    Mu = An({
      tan_: function (t) {
        var e = mn(t, 'x', 'tan')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.tan(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return t.div(n.cos().square())
              },
            }
          }
        )
      },
    }),
    Bu = An({
      tanh_: function (t) {
        var e = mn(t, 'x', 'tanh')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.tanh(e)
            return n([r]), r
          },
          { x: e },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return On(1).sub(n.square()).mulStrict(t)
              },
            }
          },
          'Tanh',
          {},
          null,
          [!0]
        )
      },
    })
  function Pu(t, e, n, r, o, a) {
    var i,
      s,
      u = mn(t, 'x', 'batchNorm'),
      c = mn(e, 'mean', 'batchNorm'),
      l = mn(n, 'variance', 'batchNorm')
    return (
      null != o && (i = mn(o, 'scale', 'batchNorm')),
      null != r && (s = mn(r, 'offset', 'batchNorm')),
      C(2 === u.rank, function () {
        return (
          'Error in batchNorm3D: x must be rank 3 but got rank ' + u.rank + '.'
        )
      }),
      C(2 === c.rank || 1 === c.rank, function () {
        return (
          'Error in batchNorm2D: mean must be rank 2 or rank 1 but got rank ' +
          c.rank +
          '.'
        )
      }),
      C(2 === l.rank || 1 === l.rank, function () {
        return (
          'Error in batchNorm2D: variance must be rank 2 or rank 1 but got rank ' +
          l.rank +
          '.'
        )
      }),
      null != i &&
        C(2 === i.rank || 1 === i.rank, function () {
          return (
            'Error in batchNorm2D: scale must be rank 2 or rank 1 but got rank ' +
            i.rank +
            '.'
          )
        }),
      null != s &&
        C(2 === s.rank || 1 === s.rank, function () {
          return (
            'Error in batchNorm2D: offset must be rank 2 or rank 1 but got rank ' +
            s.rank +
            '.'
          )
        }),
      Uu(u, c, l, s, i, a)
    )
  }
  function Lu(t, e, n, r, o, a) {
    var i,
      s,
      u = mn(t, 'x', 'batchNorm'),
      c = mn(e, 'mean', 'batchNorm'),
      l = mn(n, 'variance', 'batchNorm')
    return (
      null != o && (i = mn(o, 'scale', 'batchNorm')),
      null != r && (s = mn(r, 'offset', 'batchNorm')),
      C(3 === u.rank, function () {
        return (
          'Error in batchNorm3D: x must be rank 3 but got rank ' + u.rank + '.'
        )
      }),
      C(3 === c.rank || 1 === c.rank, function () {
        return (
          'Error in batchNorm3D: mean must be rank 3 or rank 1 but got rank ' +
          c.rank +
          '.'
        )
      }),
      C(3 === l.rank || 1 === l.rank, function () {
        return (
          'Error in batchNorm3D: variance must be rank 3 or rank 1 but got rank ' +
          l.rank +
          '.'
        )
      }),
      null != i &&
        C(3 === i.rank || 1 === i.rank, function () {
          return (
            'Error in batchNorm3D: scale must be rank 3 or rank 1 but got rank ' +
            i.rank +
            '.'
          )
        }),
      null != s &&
        C(3 === s.rank || 1 === s.rank, function () {
          return (
            'Error in batchNorm3D: offset must be rank 3 or rank 1 but got rank ' +
            s.rank +
            '.'
          )
        }),
      Uu(u, c, l, s, i, a)
    )
  }
  function Wu(t, e, n, r, o, a) {
    var i,
      s,
      u = mn(t, 'x', 'batchNorm'),
      c = mn(e, 'mean', 'batchNorm'),
      l = mn(n, 'variance', 'batchNorm')
    return (
      null != o && (i = mn(o, 'scale', 'batchNorm')),
      null != r && (s = mn(r, 'offset', 'batchNorm')),
      C(4 === u.rank, function () {
        return (
          'Error in batchNorm4D: x must be rank 4 but got rank ' + u.rank + '.'
        )
      }),
      C(4 === c.rank || 1 === c.rank, function () {
        return (
          'Error in batchNorm4D: mean must be rank 4 or rank 1 but got rank ' +
          c.rank +
          '.'
        )
      }),
      C(4 === l.rank || 1 === l.rank, function () {
        return (
          'Error in batchNorm4D: variance must be rank 4 or rank 1 but got rank ' +
          l.rank +
          '.'
        )
      }),
      null != i &&
        C(4 === i.rank || 1 === i.rank, function () {
          return (
            'Error in batchNorm4D: scale must be rank 4 or rank 1 but got rank ' +
            i.rank +
            '.'
          )
        }),
      null != s &&
        C(4 === s.rank || 1 === s.rank, function () {
          return (
            'Error in batchNorm4D: offset must be rank 4 or rank 1 but got rank ' +
            s.rank +
            '.'
          )
        }),
      Uu(u, c, l, s, i, a)
    )
  }
  function Uu(t, e, n, r, o, a) {
    null == a && (a = 0.001)
    var i,
      s,
      u,
      c = mn(t, 'x', 'batchNorm'),
      l = mn(e, 'mean', 'batchNorm'),
      h = mn(n, 'variance', 'batchNorm')
    null != o && (i = mn(o, 'scale', 'batchNorm')),
      null != r && (s = mn(r, 'offset', 'batchNorm')),
      C(l.rank === h.rank, function () {
        return 'Batch normalization gradient requires mean and variance to have equal ranks.'
      }),
      C(null == s || l.rank === s.rank, function () {
        return 'Batch normalization gradient requires mean and offset to have equal ranks.'
      }),
      C(null == i || l.rank === i.rank, function () {
        return 'Batch normalization gradient requires mean and scale to have equal ranks.'
      }),
      (u =
        0 === c.rank || 1 === c.rank
          ? c.as4D(1, 1, 1, c.size)
          : 2 === c.rank
          ? c.as4D(1, 1, c.shape[0], c.shape[1])
          : 3 === c.rank
          ? c.as4D(1, c.shape[0], c.shape[1], c.shape[2])
          : c)
    var f = [c, l, h, i]
    return Lt.runKernelFunc(
      function (t, e) {
        var n = t.batchNormalization(u, Vu(l), Vu(h), a, Vu(i), Vu(s))
        return e([c, l, h, i]), n
      },
      { x: c, mean: l, variance: h, scale: i, offset: s },
      function (t, e) {
        var n = e,
          r = n[0],
          o = n[1],
          i = n[2],
          s = n[3],
          c = null == s ? On(1) : s,
          l = Eo(o.shape, u.shape),
          h = []
        if (1 === o.rank) {
          for (var f = 0; f < u.shape.length - 1; ++f) h.push(u.shape[f])
          h.push(1)
        }
        var d = r.sub(o),
          p = t.mul(c),
          v = Ru(i.add(On(a))),
          m = v.mul(v).mul(v).mul(On(-0.5))
        return {
          x: function () {
            return 1 === o.rank
              ? t
                  .mul(Lr(v.as4D(1, 1, 1, o.shape[0]), h))
                  .mul(c)
                  .reshape(r.shape)
              : t.mul(v).mul(c).reshape(r.shape)
          },
          mean: function () {
            var t = v.mul(On(-1)).mul(p)
            return 1 === o.rank && (t = t.sum(l)), t.reshape(o.shape)
          },
          variance: function () {
            var t = m.mul(d).mul(p)
            return 1 === o.rank && (t = t.sum(l)), t.reshape(o.shape)
          },
          scale: function () {
            var e = d.mul(v),
              n = t.mul(e)
            return 1 === o.rank && (n = n.sum(l)), n.reshape(o.shape)
          },
          offset: function () {
            var e = t
            return 1 === o.rank && (e = e.sum(l)), e.reshape(o.shape)
          },
        }
      },
      'BatchNormalization',
      { varianceEpsilon: a },
      f
    ).reshape(c.shape)
  }
  function Vu(t) {
    return null == t
      ? null
      : 0 === t.rank
      ? t.as1D()
      : 1 === t.rank
      ? t
      : 2 === t.rank
      ? t.as4D(1, 1, t.shape[0], t.shape[1])
      : 3 === t.rank
      ? t.as4D(1, t.shape[0], t.shape[1], t.shape[2])
      : t
  }
  function zu() {
    Xe(
      'tf.batchNormalization() is going away. Use tf.batchNorm() instead, and note the positional argument change of scale, offset, and varianceEpsilon'
    )
  }
  var Gu = An({
      batchNormalization2d_: function (t, e, n, r, o, a) {
        return void 0 === r && (r = 0.001), zu(), Pu(t, e, n, a, o, r)
      },
    }),
    Hu = An({
      batchNormalization3d_: function (t, e, n, r, o, a) {
        return void 0 === r && (r = 0.001), zu(), Lu(t, e, n, a, o, r)
      },
    }),
    qu = An({
      batchNormalization4d_: function (t, e, n, r, o, a) {
        return void 0 === r && (r = 0.001), zu(), Wu(t, e, n, a, o, r)
      },
    }),
    Ku = An({
      batchNormalization_: function (t, e, n, r, o, a) {
        return void 0 === r && (r = 0.001), zu(), Uu(t, e, n, a, o, r)
      },
    }),
    ju = An({ batchNorm_: Uu }),
    Xu = An({ batchNorm2d_: Pu }),
    Yu = An({ batchNorm3d_: Lu }),
    $u = An({ batchNorm4d_: Wu })
  var Qu = An({
      logicalAnd_: function (t, e) {
        var n = mn(t, 'a', 'logicalAnd', 'bool'),
          r = mn(e, 'b', 'logicalAnd', 'bool')
        return (
          Ro(n.shape, r.shape),
          Lt.runKernelFunc(
            function (t) {
              return t.logicalAnd(n, r)
            },
            { a: n, b: r },
            null,
            'LogicalAnd'
          )
        )
      },
    }),
    Ju = An({
      logicalNot_: function (t) {
        var e = mn(t, 'x', 'logicalNot', 'bool')
        return Lt.runKernelFunc(
          function (t) {
            return t.logicalNot(e)
          },
          { $x: e }
        )
      },
    }),
    Zu = An({
      logicalOr_: function (t, e) {
        var n = mn(t, 'a', 'logicalOr', 'bool'),
          r = mn(e, 'b', 'logicalOr', 'bool')
        return (
          Ro(n.shape, r.shape),
          Lt.runKernelFunc(
            function (t) {
              return t.logicalOr(n, r)
            },
            { $a: n, $b: r }
          )
        )
      },
    }),
    tc = An({
      logicalXor_: function (t, e) {
        var n = mn(t, 'a', 'logicalXor', 'bool'),
          r = mn(e, 'b', 'logicalXor', 'bool')
        return Ro(n.shape, r.shape), Zu(t, e).logicalAnd(Qu(t, e).logicalNot())
      },
    }),
    ec = An({
      where_: function (t, e, n) {
        var r = mn(e, 'a', 'where'),
          o = mn(n, 'b', 'where'),
          a = mn(t, 'condition', 'where', 'bool')
        return (
          E(r.shape, o.shape, 'Error in where: '),
          1 === a.rank
            ? C(a.shape[0] === r.shape[0], function () {
                return 'The first dimension of `a` must match the size of `condition`.'
              })
            : E(a.shape, o.shape, 'Error in where: '),
          Lt.runKernelFunc(
            function (t, e) {
              var n = t.select(a, r, o)
              return e([a]), n
            },
            { $condition: a, $a: r, $b: o },
            function (t, e) {
              var n = e[0]
              return {
                $condition: function () {
                  return Xn(n).toFloat()
                },
                $a: function () {
                  return t.mul(n.cast(t.dtype))
                },
                $b: function () {
                  return t.mul(n.logicalNot().cast(t.dtype))
                },
              }
            }
          )
        )
      },
    }),
    nc = function (t) {
      return n(this, void 0, void 0, function () {
        var e, n, o
        return r(this, function (r) {
          switch (r.label) {
            case 0:
              return [4, (e = mn(t, 'condition', 'whereAsync', 'bool')).data()]
            case 1:
              return (
                (n = r.sent()),
                (o = na(e.shape, n)),
                t !== e && e.dispose(),
                [2, o]
              )
          }
        })
      })
    }
  var rc = An({
      add_: function (t, e) {
        var n,
          r = mn(t, 'a', 'add'),
          o = mn(e, 'b', 'add')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t) {
            return t.add(r, o)
          },
          { a: r, b: o },
          function (t) {
            return {
              a: function () {
                var e = t,
                  n = Eo(r.shape, a)
                return n.length > 0 && (e = e.sum(n)), e.reshape(r.shape)
              },
              b: function () {
                var e = t,
                  n = Eo(o.shape, a)
                return n.length > 0 && (e = e.sum(n)), e.reshape(o.shape)
              },
            }
          },
          'Add'
        )
      },
    }),
    oc = An({
      addN_: function (t) {
        C(Array.isArray(t), function () {
          return 'The argument passed to tf.addN() must be a list of tensors'
        }),
          C(t.length >= 1, function () {
            return (
              'Must pass at least one tensor to tf.addN(), but got ' + t.length
            )
          })
        var e = t.map(function (t, e) {
            return mn(t, 'tensors' + e, 'addN')
          }),
          n = e[0]
        e.forEach(function (t) {
          if (t.dtype !== n.dtype)
            throw new Error(
              'All tensors passed to tf.addN() must have the same dtype'
            )
        }),
          e.forEach(function (t) {
            if (!S(t.shape, n.shape))
              throw new Error(
                'All tensors passed to tf.addN() must have the same shape'
              )
          })
        var r = e
        return Lt.runKernelFunc(
          function (t) {
            return t.addN(e)
          },
          r,
          function (t) {
            var n = {}
            return (
              e.forEach(function (e, r) {
                n[r] = function () {
                  return t.clone()
                }
              }),
              n
            )
          },
          'AddN'
        )
      },
    }),
    ac = An({
      addStrict_: function (t, e) {
        var n = mn(t, 'a', 'addStrict'),
          r = mn(e, 'b', 'addStrict')
        return E(n.shape, r.shape, 'Error in addStrict: '), n.add(r)
      },
    }),
    ic = An({
      atan2_: function (t, e) {
        var n,
          r = mn(t, 'a', 'atan2'),
          o = mn(e, 'b', 'atan2')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.atan2(r, o)
            return e([r, o]), n
          },
          { $a: r, $b: o },
          function (t, e) {
            var n = e[0],
              r = e[1]
            return {
              $a: function () {
                var e = rc(n.square(), r.square()),
                  o = t.mul(r.div(e)),
                  i = Eo(n.shape, a)
                return i.length > 0 && (o = o.sum(i)), o.reshape(n.shape)
              },
              $b: function () {
                var e = rc(n.square(), r.square()),
                  o = wu(t.mul(n.div(e))),
                  i = Eo(r.shape, a)
                return i.length > 0 && (o = o.sum(i)), o.reshape(r.shape)
              },
            }
          }
        )
      },
    }),
    sc = An({
      div_: function (t, e) {
        var n,
          r = mn(t, 'a', 'div'),
          o = mn(e, 'b', 'div')
        if (
          ((n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          'int32' === r.dtype && 'int32' === o.dtype)
        )
          return lc(r, o)
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.realDivide(r, o)
            return e([r, o]), n
          },
          { a: r, b: o },
          function (t, e) {
            var n = e[0],
              r = e[1]
            return {
              a: function () {
                var e = t.div(r.toFloat()),
                  o = Eo(n.shape, a)
                return o.length > 0 ? e.sum(o).reshape(n.shape) : e
              },
              b: function () {
                var e = t.mul(n.toFloat()),
                  o = Eo(r.shape, a)
                o.length > 0 && (e = e.sum(o).reshape(r.shape))
                var i = r.square()
                return e.div(i.toFloat()).neg()
              },
            }
          },
          'Div'
        )
      },
    }),
    uc = An({
      divNoNan_: function (t, e) {
        var n,
          r = mn(t, 'a', 'div'),
          o = mn(e, 'b', 'div')
        ;(r = (n = Nt(r, o))[0]), (o = n[1])
        var a = sc(r, o),
          i = Xn(a),
          s = o.equal(i)
        return ec(s, i, a)
      },
    }),
    cc = An({
      divStrict_: function (t, e) {
        var n = mn(t, 'a', 'div'),
          r = mn(e, 'b', 'div')
        return E(n.shape, r.shape, 'Error in divideStrict: '), n.div(r)
      },
    }),
    lc = An({
      floorDiv_: function (t, e) {
        var n,
          r = mn(t, 'a', 'floorDiv'),
          o = mn(e, 'b', 'floorDiv')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.floorDiv(r, o)
            return e([r, o]), n
          },
          { a: r, b: o },
          function (t, e) {
            var n = e[0],
              r = e[1]
            return {
              a: function () {
                var e = t.div(r.toFloat()),
                  o = Eo(n.shape, a)
                return o.length > 0 ? e.sum(o).reshape(n.shape) : e
              },
              b: function () {
                var e = t.mul(n.toFloat()),
                  o = Eo(r.shape, a)
                o.length > 0 && (e = e.sum(o).reshape(r.shape))
                var i = r.square()
                return e.div(i.toFloat()).neg()
              },
            }
          },
          'FloorDiv'
        )
      },
    }),
    hc = An({
      maximum_: function (t, e) {
        var n,
          r = mn(t, 'a', 'maximum'),
          o = mn(e, 'b', 'maximum')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          'bool' === r.dtype && ((r = r.toInt()), (o = o.toInt())),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t, e) {
              var n = t.maximum(r, o)
              return e([r, o]), n
            },
            { a: r, b: o },
            function (t, e) {
              var n = e[0],
                r = e[1]
              return {
                a: function () {
                  return t.mul(n.greaterEqual(r).toFloat())
                },
                b: function () {
                  return t.mul(n.less(r).toFloat())
                },
              }
            },
            'Maximum'
          )
        )
      },
    }),
    fc = An({
      maximumStrict_: function (t, e) {
        var n = mn(t, 'a', 'maximumStrict'),
          r = mn(e, 'b', 'maximumStrict')
        return E(n.shape, r.shape, 'Error in maximumStrict: '), n.maximum(r)
      },
    }),
    dc = An({
      minimum_: function (t, e) {
        var n,
          r = mn(t, 'a', 'minimum'),
          o = mn(e, 'b', 'minimum')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          'bool' === r.dtype && ((r = r.toInt()), (o = o.toInt())),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t, e) {
              var n = t.minimum(r, o)
              return e([r, o]), n
            },
            { a: r, b: o },
            function (t, e) {
              var n = e[0],
                r = e[1]
              return {
                a: function () {
                  return t.mul(n.lessEqual(r).toFloat())
                },
                b: function () {
                  return t.mul(n.greater(r).toFloat())
                },
              }
            },
            'Minimum'
          )
        )
      },
    }),
    pc = An({
      minimumStrict_: function (t, e) {
        var n = mn(t, 'a', 'minimumStrict'),
          r = mn(e, 'b', 'minimumStrict')
        return E(n.shape, r.shape, 'Error in minimumStrict: '), n.minimum(r)
      },
    }),
    vc = An({
      mod_: function (t, e) {
        var n,
          r = mn(t, 'a', 'mod'),
          o = mn(e, 'b', 'mod')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.mod(r, o)
            return e([r, o]), n
          },
          { $a: r, $b: o },
          function (t, e) {
            var n = e[0],
              r = e[1]
            return {
              $a: function () {
                var e = Eo(n.shape, a)
                return e.length > 0 ? t.sum(e).reshape(n.shape) : t
              },
              $b: function () {
                var e = t.mul(n.div(r).floor().neg()),
                  o = Eo(r.shape, a)
                return o.length > 0 ? e.sum(o).reshape(r.shape) : e
              },
            }
          }
        )
      },
    }),
    mc = An({
      modStrict_: function (t, e) {
        var n = mn(t, 'a', 'modStrict'),
          r = mn(e, 'b', 'modStrict')
        return E(n.shape, r.shape, 'Error in modStrict: '), n.mod(r)
      },
    }),
    gc = An({
      mul_: function (t, e) {
        var n,
          r = mn(t, 'a', 'mul'),
          o = mn(e, 'b', 'mul')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.multiply(r, o)
            return e([r, o]), n
          },
          { a: r, b: o },
          function (t, e) {
            var n = e[0],
              r = e[1]
            return {
              a: function () {
                var e = t.mul(r.toFloat()),
                  o = Eo(n.shape, a)
                return o.length > 0 ? e.sum(o).reshape(n.shape) : e
              },
              b: function () {
                var e = t.mul(n.toFloat()),
                  o = Eo(r.shape, a)
                return o.length > 0 ? e.sum(o).reshape(r.shape) : e
              },
            }
          },
          'Mul'
        )
      },
    }),
    yc = An({
      mulStrict_: function (t, e) {
        var n = mn(t, 'a', 'mul'),
          r = mn(e, 'b', 'mul')
        return E(n.shape, r.shape, 'Error in multiplyStrict: '), n.mul(r)
      },
    }),
    xc = An({
      pow_: function (t, e) {
        var n,
          r = mn(t, 'base', 'pow'),
          o = mn(e, 'exp', 'pow')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape),
          i = [r, o]
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.pow(r, o)
            return e([r, o, n]), n
          },
          { a: r, b: o },
          function (t, e) {
            var n = e[0],
              r = e[1],
              o = e[2]
            return {
              a: function () {
                var e = r.toFloat(),
                  o = t.mul(e.mul(n.pow(e.sub(On(1))))),
                  i = Eo(n.shape, a)
                return i.length > 0 && (o = o.sum(i)), o.reshape(n.shape)
              },
              b: function () {
                var e = n.greater(0),
                  i = n.log().where(e, Xn(n)),
                  s = t.mul(o.mul(i)),
                  u = Eo(r.shape, a)
                return u.length > 0 && (s = s.sum(u)), s.reshape(r.shape)
              },
            }
          },
          'Pow',
          {},
          i,
          [!0]
        )
      },
    }),
    bc = An({
      powStrict_: function (t, e) {
        return E(t.shape, e.shape, 'Error in powStrict: '), t.pow(e)
      },
    }),
    wc = An({
      squaredDifferenceStrict_: function (t, e) {
        var n = mn(t, 'a', 'squaredDifferenceStrict'),
          r = mn(e, 'b', 'squaredDifferenceStrict')
        return (
          E(n.shape, r.shape, 'Error in squaredDifferenceStrict: '),
          n.squaredDifference(r)
        )
      },
    }),
    Cc = An({
      sub_: function (t, e) {
        var n,
          r = mn(t, 'a', 'sub'),
          o = mn(e, 'b', 'sub')
        ;(n = Nt(r, o)), (r = n[0]), (o = n[1])
        var a = Ro(r.shape, o.shape)
        return Lt.runKernelFunc(
          function (t) {
            return t.subtract(r, o)
          },
          { a: r, b: o },
          function (t) {
            return {
              a: function () {
                var e = t,
                  n = Eo(r.shape, a)
                return n.length > 0 && (e = e.sum(n)), e.reshape(r.shape)
              },
              b: function () {
                var e = t,
                  n = Eo(o.shape, a)
                return n.length > 0 && (e = e.sum(n)), e.neg().reshape(o.shape)
              },
            }
          },
          'Sub'
        )
      },
    }),
    Ec = An({
      subStrict_: function (t, e) {
        var n = mn(t, 'a', 'subStrict'),
          r = mn(e, 'b', 'subStrict')
        return E(n.shape, r.shape, 'Error in subStrict: '), n.sub(r)
      },
    })
  var Rc = An({
      equal_: function (t, e) {
        var n,
          r = mn(t, 'a', 'equal'),
          o = mn(e, 'b', 'equal')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t) {
              return t.equal(r, o)
            },
            { $a: r, $b: o }
          )
        )
      },
    }),
    Ic = An({
      equalStrict_: function (t, e) {
        var n = mn(t, 'a', 'equalStrict'),
          r = mn(e, 'b', 'equalStrict')
        return E(n.shape, r.shape, 'Error in equalStrict: '), n.equal(r)
      },
    }),
    kc = An({
      greater_: function (t, e) {
        var n,
          r = mn(t, 'a', 'greater'),
          o = mn(e, 'b', 'greater')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t) {
              return t.greater(r, o)
            },
            { a: r, b: o },
            null,
            'Greater'
          )
        )
      },
    }),
    Sc = An({
      greaterEqual_: function (t, e) {
        var n,
          r = mn(t, 'a', 'greaterEqual'),
          o = mn(e, 'b', 'greaterEqual')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t, e) {
              var n = t.greaterEqual(r, o)
              return e([r, o]), n
            },
            { a: r, b: o },
            function (t, e) {
              var n = e[0],
                r = e[1]
              return {
                a: function () {
                  return Xn(n)
                },
                b: function () {
                  return Xn(r)
                },
              }
            },
            'GreaterEqual'
          )
        )
      },
    }),
    Ac = An({
      greaterEqualStrict_: function (t, e) {
        var n = mn(t, 'a', 'greaterEqualStrict'),
          r = mn(e, 'b', 'greaterEqualStrict')
        return (
          E(n.shape, r.shape, 'Error in greaterEqualStrict: '),
          n.greaterEqual(r)
        )
      },
    }),
    Dc = An({
      greaterStrict_: function (t, e) {
        var n = mn(t, 'a', 'greaterStrict'),
          r = mn(e, 'b', 'greaterStrict')
        return E(n.shape, r.shape, 'Error in greaterStrict: '), n.greater(r)
      },
    }),
    Tc = An({
      less_: function (t, e) {
        var n,
          r = mn(t, 'a', 'less'),
          o = mn(e, 'b', 'less')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t) {
              return t.less(r, o)
            },
            { a: r, b: o },
            null,
            'Less'
          )
        )
      },
    }),
    Nc = An({
      lessEqual_: function (t, e) {
        var n,
          r = mn(t, 'a', 'lessEqual'),
          o = mn(e, 'b', 'lessEqual')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t, e) {
              var n = t.lessEqual(r, o)
              return e([r, o]), n
            },
            { a: r, b: o },
            null,
            'LessEqual'
          )
        )
      },
    }),
    Fc = An({
      lessEqualStrict_: function (t, e) {
        var n = mn(t, 'a', 'lessEqualStrict'),
          r = mn(e, 'b', 'lessEqualStrict')
        return E(n.shape, r.shape, 'Error in lessEqualStrict: '), n.lessEqual(r)
      },
    }),
    _c = An({
      lessStrict_: function (t, e) {
        var n = mn(t, 'a', 'lessStrict'),
          r = mn(e, 'b', 'lessStrict')
        return E(n.shape, r.shape, 'Error in lessStrict: '), n.less(r)
      },
    }),
    Oc = An({
      notEqual_: function (t, e) {
        var n,
          r = mn(t, 'a', 'notEqual'),
          o = mn(e, 'b', 'notEqual')
        return (
          (n = Nt(r, o)),
          (r = n[0]),
          (o = n[1]),
          Ro(r.shape, o.shape),
          Lt.runKernelFunc(
            function (t) {
              return t.notEqual(r, o)
            },
            { a: r, b: o },
            null,
            'NotEqual'
          )
        )
      },
    }),
    Mc = An({
      notEqualStrict_: function (t, e) {
        var n = mn(t, 'a', 'notEqualStrict'),
          r = mn(e, 'b', 'notEqualStrict')
        return E(n.shape, r.shape, 'Error in notEqualStrict: '), n.notEqual(r)
      },
    })
  function Bc(t, e) {
    for (var n = [], r = t; r < e; ++r) n.push(r)
    return n
  }
  function Pc(t) {
    for (var e = [], n = 0; n < t.length; ++n)
      for (var r = 0; r < t[n].length; ++r) e.push(t[n][r])
    return e
  }
  var Lc = An({
      gather_: function (t, e, n) {
        void 0 === n && (n = 0)
        var r = mn(t, 'x', 'gather'),
          o = mn(e, 'indices', 'gather', 'int32')
        n = O(n, r.shape)[0]
        var a = (function (t, e, n) {
          for (var r = t.shape[n], o = [], a = 1, i = 1, s = 0; s < n; s++)
            o.push(t.shape[s]), (a *= t.shape[s])
          for (s = 0; s < e.rank; s++) o.push(e.shape[s])
          for (s = n + 1; s < t.rank; s++) o.push(t.shape[s]), (i *= t.shape[s])
          return { batchSize: a, sliceSize: i, dimSize: r, outputShape: o }
        })(r, o, n)
        return Lt.runKernelFunc(
          function (t, e) {
            var a = t.gather(r, o.flatten(), n)
            return e([o]), a
          },
          { x: r, indices: o },
          function (t, e) {
            var o = e[0]
            return {
              x: function () {
                var e = r.shape,
                  a = o.size,
                  i = e.slice(0, n),
                  s = i.length,
                  u = e.slice(n, e.length).slice(1),
                  c = u.length,
                  l = Bc(0, s),
                  h = Bc(s + 1, s + 1 + c),
                  f = Pc([i, [a], u]),
                  d = t.reshape(f),
                  p = o.reshape([a]),
                  v = Pc([[s], l, h]),
                  m = d.transpose(v),
                  g = Wc(m, p, r.shape[n]),
                  y = Rn(v)
                return (g = g.transpose(y))
              },
              indices: function () {
                return o
              },
            }
          },
          'Gather',
          { axis: n }
        ).reshape(a.outputShape)
      },
    }),
    Wc = An({
      unsortedSegmentSum_: function (t, e, n) {
        var r = mn(t, 'x', 'unsortedSegmentSum'),
          o = mn(e, 'segmentIds', 'unsortedSegmentSum', 'int32')
        return (
          C(A(n), function () {
            return 'numSegments must be of dtype int'
          }),
          Lt.runKernelFunc(
            function (t, e) {
              var a = t.unsortedSegmentSum(r, o, n)
              return e([o]), a
            },
            { $x: r },
            function (t, e) {
              var n = e[0]
              return {
                $x: function () {
                  return (function (t, e) {
                    for (
                      var n = hc(e, Xn(e)),
                        r = Lc(t, n),
                        o = Sc(e, On(0, 'int32')),
                        a = r.rank - o.rank,
                        i = 0;
                      i < a;
                      ++i
                    )
                      o = wr(o, i + 1)
                    o = Qu(o, zn(r.shape, 'bool'))
                    var s = Xn(r)
                    return ec(o, r, s)
                  })(t, n)
                },
              }
            }
          )
        )
      },
    })
  var Uc = function (t, e, o) {
    return n(this, void 0, void 0, function () {
      var n, a, i, s, u, c, l, h, f, d, p, v, m
      return r(this, function (r) {
        switch (r.label) {
          case 0:
            for (
              n = mn(t, 'tensor', 'boolMask'),
                a = mn(e, 'mask', 'boolMask', 'bool'),
                i = null == o ? 0 : o,
                s = a.rank,
                u = n.shape,
                C(s > 0, function () {
                  return 'mask cannot be scalar'
                }),
                E(
                  u.slice(i, i + s),
                  a.shape,
                  "mask's shape must match the first K dimensions of tensor's shape,"
                ),
                c = 1,
                l = i;
              l < i + s;
              l++
            )
              c *= u[l]
            return (
              (h = u.slice(0, i).concat([c], u.slice(i + s))),
              (f = n.reshape(h)),
              (d = a.reshape([-1])),
              [4, nc(d)]
            )
          case 1:
            return (
              (p = r.sent()),
              (v = p.squeeze([1])),
              (m = Lc(f, v, i)),
              t !== n && n.dispose(),
              e !== a && a.dispose(),
              v.dispose(),
              f.dispose(),
              d.dispose(),
              p.dispose(),
              [2, m]
            )
        }
      })
    })
  }
  function Vc(t, e, n, r, o, a, i) {
    void 0 === a && (a = 'NHWC'),
      C(t.length === e.rank, function () {
        return (
          'Length of inShape (' +
          t.length +
          ') and rank of dy (' +
          e.rank +
          ') must match'
        )
      })
    var s = t,
      u = e,
      c = !1
    3 === e.rank &&
      ((c = !0),
      (u = e.as4D(1, e.shape[0], e.shape[1], e.shape[2])),
      (s = [1, t[0], t[1], t[2]])),
      C(4 === s.length, function () {
        return (
          'Error in conv2dDerInput: inShape must be length 4, but got length ' +
          s.length +
          '.'
        )
      }),
      C(4 === u.rank, function () {
        return (
          'Error in conv2dDerInput: dy must be rank 4, but got rank ' + u.rank
        )
      }),
      C(4 === n.rank, function () {
        return (
          'Error in conv2dDerInput: filter must be rank 4, but got rank ' +
          n.rank
        )
      })
    var l = 'NHWC' === a ? s[3] : s[1],
      h = 'NHWC' === a ? u.shape[3] : u.shape[1]
    C(l === n.shape[2], function () {
      return (
        'Error in conv2dDerInput: depth of input (' +
        l +
        ') must match input depth for filter ' +
        n.shape[2] +
        '.'
      )
    }),
      C(h === n.shape[3], function () {
        return (
          'Error in conv2dDerInput: depth of output (' +
          h +
          ') must match output depth for filter ' +
          n.shape[3] +
          '.'
        )
      }),
      null != i &&
        C(A(o), function () {
          return (
            'Error in conv2dDerInput: pad must be an integer when using, dimRoundingMode ' +
            i +
            ' but got pad ' +
            o +
            '.'
          )
        })
    var f = Bo(a),
      d = So(s, n.shape, r, 1, o, i, !1, f),
      p = Lt.runKernelFunc(
        function (t, e) {
          var r = t.conv2dDerInput(u, n, d)
          return e([n, u]), r
        },
        { dy4D: u, filter: n },
        function (t, e) {
          var n = e[0],
            s = e[1]
          return {
            dy4D: function () {
              return qc(t, n, r, o, a, 1, i)
            },
            filter: function () {
              return jc(t, s, n.shape, r, o, a, i)
            },
          }
        }
      )
    return c ? p.as3D(p.shape[1], p.shape[2], p.shape[3]) : p
  }
  function zc(t) {
    var e = (function (t) {
        return 'number' == typeof t
          ? [t, t, t]
          : 2 === t.length
          ? [t[0], t[1], 1]
          : t
      })(t),
      n = e[0],
      r = e[1],
      o = e[2]
    return 1 === n && 1 === r && 1 === o
  }
  function Gc(t, e, n, r, o) {
    C(t.length === e.rank, function () {
      return (
        'Length of inShape (' +
        t.length +
        ') and rank of dy (' +
        e.rank +
        ') must match'
      )
    })
    var a = t,
      i = e,
      s = !1
    4 === e.rank &&
      ((s = !0),
      (i = e.as5D(1, e.shape[0], e.shape[1], e.shape[2], e.shape[3])),
      (a = [1, t[0], t[1], t[2], t[3]]))
    var u = a[4],
      c = i.shape[4]
    C(5 === a.length, function () {
      return (
        'Error in conv3dDerInput: inShape must be length 5, but got length ' +
        a.length +
        '.'
      )
    }),
      C(5 === i.rank, function () {
        return (
          'Error in conv3dDerInput: dy must be rank 5, but got rank ' + i.rank
        )
      }),
      C(5 === n.rank, function () {
        return (
          'Error in conv3dDerInput: filter must be rank 5, but got rank ' +
          n.rank
        )
      }),
      C(u === n.shape[3], function () {
        return (
          'Error in conv3dDerInput: depth of input (' +
          u +
          ') must match input depth for filter ' +
          n.shape[3] +
          '.'
        )
      }),
      C(c === n.shape[4], function () {
        return (
          'Error in conv3dDerInput: depth of output (' +
          c +
          ') must match output depth for filter ' +
          n.shape[4] +
          '.'
        )
      })
    var l = Ao(a, n.shape, r, 1, o),
      h = Lt.runKernelFunc(
        function (t) {
          return t.conv3dDerInput(i, n, l)
        },
        { dy5D: i }
      )
    return s ? h.as4D(h.shape[1], h.shape[2], h.shape[3], h.shape[4]) : h
  }
  var Hc = An({
      conv1d_: function (t, e, n, r, o, a, i) {
        void 0 === o && (o = 'NWC'), void 0 === a && (a = 1)
        var s = mn(t, 'x', 'conv1d'),
          u = mn(e, 'filter', 'conv1d'),
          c = s,
          l = !1
        2 === s.rank && ((l = !0), (c = s.as3D(1, s.shape[0], s.shape[1]))),
          C(3 === c.rank, function () {
            return (
              'Error in conv1d: input must be rank 3, but got rank ' +
              c.rank +
              '.'
            )
          }),
          C(3 === u.rank, function () {
            return (
              'Error in conv1d: filter must be rank 3, but got rank ' +
              u.rank +
              '.'
            )
          }),
          null != i &&
            C(A(r), function () {
              return (
                'Error in conv1d: pad must be an integer when using, dimRoundingMode ' +
                i +
                ' but got pad ' +
                r +
                '.'
              )
            }),
          C(c.shape[2] === u.shape[1], function () {
            return (
              'Error in conv1d: depth of input (' +
              c.shape[2] +
              ') must match input depth for filter ' +
              u.shape[1] +
              '.'
            )
          }),
          C(Mo(n, a), function () {
            return (
              'Error in conv1D: Either stride or dilation must be 1. Got stride ' +
              n +
              " and dilation '" +
              a +
              "'"
            )
          }),
          C('NWC' === o, function () {
            return (
              'Error in conv1d: got dataFormat of ' +
              o +
              ' but only NWC is currently supported.'
            )
          })
        var h = u.as4D(1, u.shape[0], u.shape[1], u.shape[2]),
          f = c.as4D(c.shape[0], 1, c.shape[1], c.shape[2]),
          d = qc(f, h, [1, n], r, 'NHWC', [1, a], i)
        return l
          ? d.as2D(d.shape[2], d.shape[3])
          : d.as3D(d.shape[0], d.shape[2], d.shape[3])
      },
    }),
    qc = An({
      conv2d_: function (t, e, n, r, o, a, i) {
        void 0 === o && (o = 'NHWC'), void 0 === a && (a = [1, 1])
        var s = mn(t, 'x', 'conv2d'),
          u = mn(e, 'filter', 'conv2d'),
          c = s,
          l = !1
        3 === s.rank &&
          ((l = !0), (c = s.as4D(1, s.shape[0], s.shape[1], s.shape[2]))),
          C(4 === c.rank, function () {
            return (
              'Error in conv2d: input must be rank 4, but got rank ' +
              c.rank +
              '.'
            )
          }),
          C(4 === u.rank, function () {
            return (
              'Error in conv2d: filter must be rank 4, but got rank ' +
              u.rank +
              '.'
            )
          }),
          null != i &&
            C(A(r), function () {
              return (
                'Error in conv2d: pad must be an integer when using, dimRoundingMode ' +
                i +
                ' but got pad ' +
                r +
                '.'
              )
            })
        var h = 'NHWC' === o ? c.shape[3] : c.shape[1]
        C(h === u.shape[2], function () {
          return (
            'Error in conv2d: depth of input (' +
            h +
            ') must match input depth for filter ' +
            u.shape[2] +
            '.'
          )
        }),
          C(Mo(n, a), function () {
            return (
              'Error in conv2D: Either strides or dilations must be 1. Got strides ' +
              n +
              " and dilations '" +
              a +
              "'"
            )
          })
        var f = Bo(o),
          d = So(c.shape, u.shape, n, a, r, i, !1, f),
          p = [u, c],
          v = Lt.runKernelFunc(
            function (t, e) {
              var n = t.conv2d(c, u, d)
              return e([u, c]), n
            },
            { x: c, filter: u },
            function (t, e) {
              var i = e,
                s = i[0],
                u = i[1]
              return (
                C(Oo(a), function () {
                  return (
                    "Error in gradient of conv2D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '" +
                    a +
                    "'"
                  )
                }),
                {
                  x: function () {
                    return Xc(u.shape, t, s, n, r, o)
                  },
                  filter: function () {
                    return jc(u, t, s.shape, n, r, o)
                  },
                }
              )
            },
            'Conv2D',
            d,
            p
          )
        return l ? v.as3D(v.shape[1], v.shape[2], v.shape[3]) : v
      },
    }),
    Kc = An({
      conv3d_: function (t, e, n, r, o, a) {
        void 0 === o && (o = 'NDHWC'), void 0 === a && (a = [1, 1, 1])
        var i = mn(t, 'x', 'conv3d'),
          s = mn(e, 'filter', 'conv3d'),
          u = i,
          c = !1
        4 === i.rank &&
          ((c = !0),
          (u = i.as5D(1, i.shape[0], i.shape[1], i.shape[2], i.shape[3]))),
          C(5 === u.rank, function () {
            return (
              'Error in conv3d: input must be rank 5, but got rank ' +
              u.rank +
              '.'
            )
          }),
          C(5 === s.rank, function () {
            return (
              'Error in conv3d: filter must be rank 5, but got rank ' +
              s.rank +
              '.'
            )
          }),
          C(u.shape[4] === s.shape[3], function () {
            return (
              'Error in conv3d: depth of input (' +
              u.shape[4] +
              ') must match input depth for filter ' +
              s.shape[3] +
              '.'
            )
          }),
          C(
            (function (t, e) {
              return zc(t) || zc(e)
            })(n, a),
            function () {
              return (
                'Error in conv3D: Either strides or dilations must be 1. Got strides ' +
                n +
                " and dilations '" +
                a +
                "'"
              )
            }
          ),
          C('NDHWC' === o, function () {
            return (
              'Error in conv3d: got dataFormat of ' +
              o +
              ' but only NDHWC is currently supported.'
            )
          })
        var l = Ao(u.shape, s.shape, n, a, r),
          h = Lt.runKernelFunc(
            function (t, e) {
              var n = t.conv3d(u, s, l)
              return e([u, s]), n
            },
            { x: u, $filter: s },
            function (t, e) {
              C(zc(a), function () {
                return (
                  "Error in gradient of conv3D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '" +
                  a +
                  "'"
                )
              })
              var o = e[0],
                i = e[1]
              return {
                x: function () {
                  return Gc(o.shape, t, i, n, r)
                },
                $filter: function () {
                  return (function (t, e, n, r, o) {
                    var a = t
                    4 === t.rank &&
                      (a = t.as5D(
                        1,
                        t.shape[0],
                        t.shape[1],
                        t.shape[2],
                        t.shape[3]
                      ))
                    var i = e
                    4 === i.rank &&
                      (i = e.as5D(
                        1,
                        e.shape[0],
                        e.shape[1],
                        e.shape[2],
                        e.shape[3]
                      ))
                    C(5 === a.rank, function () {
                      return (
                        'Error in conv3dDerFilter: input must be rank 5, but got shape ' +
                        a.shape +
                        '.'
                      )
                    }),
                      C(5 === i.rank, function () {
                        return (
                          'Error in conv3dDerFilter: dy must be rank 5, but got shape ' +
                          i.shape +
                          '.'
                        )
                      }),
                      C(5 === n.length, function () {
                        return (
                          'Error in conv3dDerFilter: filterShape must be length 5, but got ' +
                          n +
                          '.'
                        )
                      }),
                      C(a.shape[4] === n[3], function () {
                        return (
                          'Error in conv3dDerFilter: depth of input ' +
                          a.shape[4] +
                          ') must match input depth in filter (' +
                          n[3] +
                          '.'
                        )
                      }),
                      C(i.shape[4] === n[4], function () {
                        return (
                          'Error in conv3dDerFilter: depth of dy (' +
                          i.shape[4] +
                          ') must match output depth for filter (' +
                          n[4] +
                          ').'
                        )
                      })
                    var s = Ao(a.shape, n, r, 1, o)
                    return Lt.runKernelFunc(
                      function (t) {
                        return t.conv3dDerFilter(a, i, s)
                      },
                      { x5D: a, dy5D: i }
                    )
                  })(o, t, i.shape, n, r)
                },
              }
            }
          )
        return c ? h.as4D(h.shape[1], h.shape[2], h.shape[3], h.shape[4]) : h
      },
    }),
    jc = An({
      conv2dDerFilter_: function (t, e, n, r, o, a, i) {
        void 0 === a && (a = 'NHWC')
        var s = t
        3 === t.rank && (s = t.as4D(1, t.shape[0], t.shape[1], t.shape[2]))
        var u = e
        3 === u.rank && (u = e.as4D(1, e.shape[0], e.shape[1], e.shape[2])),
          C(4 === s.rank, function () {
            return (
              'Error in conv2dDerFilter: input must be rank 4, but got shape ' +
              s.shape +
              '.'
            )
          }),
          C(4 === u.rank, function () {
            return (
              'Error in conv2dDerFilter: dy must be rank 4, but got shape ' +
              u.shape +
              '.'
            )
          }),
          C(4 === n.length, function () {
            return (
              'Error in conv2dDerFilter: filterShape must be length 4, but got ' +
              n +
              '.'
            )
          })
        var c = 'NHWC' === a ? s.shape[3] : s.shape[1],
          l = 'NHWC' === a ? u.shape[3] : u.shape[1]
        C(c === n[2], function () {
          return (
            'Error in conv2dDerFilter: depth of input ' +
            c +
            ') must match input depth in filter (' +
            n[2] +
            '.'
          )
        }),
          C(l === n[3], function () {
            return (
              'Error in conv2dDerFilter: depth of dy (' +
              l +
              ') must match output depth for filter (' +
              n[3] +
              ').'
            )
          }),
          null != i &&
            C(A(o), function () {
              return (
                'Error in conv2dDerFilter: pad must be an integer when using, dimRoundingMode ' +
                i +
                ' but got pad ' +
                o +
                '.'
              )
            })
        var h = Bo(a),
          f = So(s.shape, n, r, 1, o, i, !1, h)
        return Lt.runKernelFunc(
          function (t) {
            return t.conv2dDerFilter(s, u, f)
          },
          { x4D: s, dy4D: u }
        )
      },
    }),
    Xc = An({ conv2dDerInput_: Vc }),
    Yc = An({
      depthwiseConv2d_: function (t, e, n, r, o, a, i) {
        void 0 === a && (a = [1, 1])
        var s = mn(t, 'x', 'depthwiseConv2d'),
          u = mn(e, 'filter', 'depthwiseConv2d'),
          c = s,
          l = !1
        3 === s.rank &&
          ((l = !0), (c = s.as4D(1, s.shape[0], s.shape[1], s.shape[2]))),
          C(4 === c.rank, function () {
            return (
              'Error in depthwiseConv2d: input must be rank 4, but got rank ' +
              c.rank +
              '.'
            )
          }),
          C(4 === u.rank, function () {
            return (
              'Error in depthwiseConv2d: filter must be rank 4, but got rank ' +
              u.rank +
              '.'
            )
          }),
          C(c.shape[3] === u.shape[2], function () {
            return (
              'Error in depthwiseConv2d: number of input channels (' +
              c.shape[3] +
              ') must match the inChannels dimension in filter ' +
              u.shape[2] +
              '.'
            )
          }),
          null == a && (a = [1, 1]),
          C(Mo(n, a), function () {
            return (
              'Error in depthwiseConv2d: Either strides or dilations must be 1. Got strides ' +
              n +
              " and dilations '" +
              a +
              "'"
            )
          }),
          null != i &&
            C(A(r), function () {
              return (
                'Error in depthwiseConv2d: pad must be an integer when using, dimRoundingMode ' +
                i +
                ' but got pad ' +
                r +
                '.'
              )
            })
        var h = So(c.shape, u.shape, n, a, r, i, !0),
          f = [c, u],
          d = Lt.runKernelFunc(
            function (t, e) {
              var n = t.depthwiseConv2D(c, u, h)
              return e([c, u]), n
            },
            { x: c, filter: u },
            function (t, e) {
              C(Oo(a), function () {
                return (
                  "Error in gradient of depthwiseConv2d: dilation rates greater than 1 are not yet supported. Got dilations '" +
                  a +
                  "'"
                )
              })
              var n = e[0],
                r = e[1]
              return {
                x: function () {
                  return $c(n.shape, t, r, h)
                },
                filter: function () {
                  return Qc(n, t, r.shape, h)
                },
              }
            },
            'DepthwiseConv2dNative',
            h,
            f
          )
        return l ? d.as3D(d.shape[1], d.shape[2], d.shape[3]) : d
      },
    }),
    $c = An({
      depthwiseConv2dDerInput_: function (t, e, n, r) {
        var o = e,
          a = !1
        3 === e.rank &&
          ((a = !0), (o = e.as4D(1, e.shape[0], e.shape[1], e.shape[2])))
        var i = Lt.runKernelFunc(
          function (t) {
            return t.depthwiseConv2DDerInput(o, n, r)
          },
          { dy4D: o }
        )
        return a ? i.as3D(i.shape[1], i.shape[2], i.shape[3]) : i
      },
    }),
    Qc = An({
      depthwiseConv2dDerFilter_: function (t, e, n, r) {
        var o = t
        3 === t.rank && (o = t.as4D(1, t.shape[0], t.shape[1], t.shape[2]))
        var a = e
        return (
          3 === a.rank && (a = e.as4D(1, e.shape[0], e.shape[1], e.shape[2])),
          Lt.runKernelFunc(
            function (t) {
              return t.depthwiseConv2DDerFilter(o, a, r)
            },
            { x4D: o, dy4D: a }
          )
        )
      },
    }),
    Jc = An({
      separableConv2d_: function (t, e, n, r, o, a, i) {
        void 0 === a && (a = [1, 1]), void 0 === i && (i = 'NHWC')
        var s = mn(t, 'x', 'separableConv2d'),
          u = mn(e, 'depthwiseFilter', 'separableConv2d'),
          c = mn(n, 'pointwiseFilter', 'separableConv2d'),
          l = s,
          h = !1
        if (
          (3 === s.rank &&
            ((h = !0), (l = s.as4D(1, s.shape[0], s.shape[1], s.shape[2]))),
          'NCHW' === i)
        )
          throw new Error(
            'separableConv2d currently does not support dataFormat NCHW; only NHWC is supported'
          )
        C(4 === l.rank, function () {
          return (
            'Error in separableConv2d: input must be rank 4, but got rank ' +
            l.rank +
            '.'
          )
        }),
          C(4 === u.rank, function () {
            return (
              'Error in separableConv2d: depthwise filter must be rank 4, but got rank ' +
              u.rank +
              '.'
            )
          }),
          C(4 === c.rank, function () {
            return (
              'Error in separableConv2d: pointwise filter must be rank 4, but got rank ' +
              u.rank +
              '.'
            )
          }),
          C(1 === c.shape[0], function () {
            return (
              'Error in separableConv2d: the first dimension of pointwise filter  must be 1, but got ' +
              c.shape[0] +
              '.'
            )
          }),
          C(1 === c.shape[1], function () {
            return (
              'Error in separableConv2d: the second dimension of pointwise filter must be 1, but got ' +
              c.shape[1] +
              '.'
            )
          })
        var f = u.shape[2],
          d = u.shape[3]
        C(c.shape[2] === f * d, function () {
          return (
            'Error in separableConv2d: the third dimension of pointwise filter must be ' +
            f * d +
            ', but got ' +
            c.shape[2] +
            '.'
          )
        })
        var p = Yc(l, u, r, o, i, a),
          v = qc(p, c, 1, 'valid', i)
        return h ? v.as3D(v.shape[1], v.shape[2], v.shape[3]) : v
      },
    }),
    Zc = An({
      conv2dTranspose_: function (t, e, n, r, o, a) {
        return Vc(
          n,
          mn(t, 'x', 'conv2dTranspose'),
          mn(e, 'filter', 'conv2dTranspose'),
          r,
          o,
          'NHWC',
          a
        )
      },
    }),
    tl = An({
      conv3dTranspose_: function (t, e, n, r, o) {
        return Gc(
          n,
          mn(t, 'x', 'conv3dTranspose'),
          mn(e, 'filter', 'conv3dTranspose'),
          r,
          o
        )
      },
    })
  var el = An({
      matMul_: function (t, e, n, r) {
        var o
        void 0 === n && (n = !1), void 0 === r && (r = !1)
        var a = mn(t, 'a', 'matMul'),
          i = mn(e, 'b', 'matMul')
        ;(o = Nt(a, i)), (a = o[0]), (i = o[1])
        var s = n ? a.shape[a.rank - 2] : a.shape[a.rank - 1],
          u = r ? i.shape[i.rank - 1] : i.shape[i.rank - 2],
          c = n ? a.shape[a.rank - 1] : a.shape[a.rank - 2],
          l = r ? i.shape[i.rank - 2] : i.shape[i.rank - 1],
          h = a.shape.slice(0, -2),
          f = i.shape.slice(0, -2),
          d = k(h),
          p = k(f)
        C(a.rank >= 2 && i.rank >= 2 && a.rank === i.rank, function () {
          return (
            'Error in matMul: inputs must have the same rank of at least 2, got ranks ' +
            a.rank +
            ' and ' +
            i.rank +
            '.'
          )
        }),
          C(S(h, f), function () {
            return (
              'Error in matMul: outer dimensions (' +
              h +
              ') and (' +
              f +
              ') of Tensors with shapes ' +
              a.shape +
              ' and ' +
              i.shape +
              ' must match.'
            )
          }),
          C(s === u, function () {
            return (
              'Error in matMul: inner shapes (' +
              s +
              ') and (' +
              u +
              ') of Tensors with shapes ' +
              a.shape +
              ' and ' +
              i.shape +
              ' and transposeA=' +
              n +
              ' and transposeB=' +
              r +
              ' must match.'
            )
          })
        var v = a.shape.slice(0, -2).concat([c, l]),
          m = n ? a.as3D(d, s, c) : a.as3D(d, c, s),
          g = r ? i.as3D(p, l, u) : i.as3D(p, u, l),
          y = { transposeA: n, transposeB: r }
        return Lt.runKernelFunc(
          function (t, e) {
            var o = t.batchMatMul(m, g, n, r)
            return e([m, g]), o
          },
          { a: m, b: g },
          function (t, e) {
            var o = e,
              a = o[0],
              i = o[1]
            return n || r
              ? !n && r
                ? {
                    a: function () {
                      return t.matMul(i, !1, !1)
                    },
                    b: function () {
                      return t.matMul(a, !0, !1)
                    },
                  }
                : n && !r
                ? {
                    a: function () {
                      return i.matMul(t, !1, !0)
                    },
                    b: function () {
                      return a.matMul(t, !1, !1)
                    },
                  }
                : {
                    a: function () {
                      return i.matMul(t, !0, !0)
                    },
                    b: function () {
                      return t.matMul(a, !0, !0)
                    },
                  }
              : {
                  a: function () {
                    return t.matMul(i, !1, !0)
                  },
                  b: function () {
                    return a.matMul(t, !0, !1)
                  },
                }
          },
          'BatchMatMul',
          y
        ).reshape(v)
      },
    }),
    nl = An({
      dot_: function (t, e) {
        var n = mn(t, 't1', 'dot'),
          r = mn(e, 't2', 'dot')
        C(
          !((1 !== n.rank && 2 !== n.rank) || (1 !== r.rank && 2 !== r.rank)),
          function () {
            return (
              'Error in dot: inputs must all be rank 1 or 2, but got ranks ' +
              n.rank +
              ' and ' +
              r.rank +
              '.'
            )
          }
        )
        var o = 1 === n.rank ? n.size : n.shape[1],
          a = 1 === r.rank ? r.size : r.shape[0]
        return (
          C(o === a, function () {
            return (
              'Error in dot: inner dimensions of inputs must match, but got ' +
              o +
              ' and ' +
              a +
              '.'
            )
          }),
          1 === n.rank && 1 === r.rank
            ? n.as2D(1, -1).matMul(r.as2D(-1, 1)).asScalar()
            : 1 === n.rank && 2 === r.rank
            ? n.as2D(1, -1).matMul(r.as2D(r.shape[0], r.shape[1])).as1D()
            : 2 === n.rank && 1 === r.rank
            ? n.matMul(r.as2D(-1, 1)).as1D()
            : n.matMul(r.as2D(r.shape[0], r.shape[1]))
        )
      },
    }),
    rl = An({
      outerProduct_: function (t, e) {
        var n = mn(t, 'v1', 'outerProduct'),
          r = mn(e, 'v2', 'outerProduct')
        return (
          C(1 === n.rank && 1 === r.rank, function () {
            return (
              'Error in outerProduct: inputs must be rank 1, but got ranks ' +
              n.rank +
              ' and ' +
              r.rank +
              '.'
            )
          }),
          n.as2D(-1, 1).matMul(r.as2D(1, -1))
        )
      },
    })
  var ol = An({
      reverse_: function (t, e) {
        var n = mn(t, 'x', 'reverse')
        if (0 === n.rank) return n.clone()
        var r = O(e, n.shape)
        return Lt.runKernelFunc(
          function (t) {
            return t.reverse(n, r)
          },
          { $x: n },
          function (t) {
            return {
              $x: function () {
                return t.reverse(r)
              },
            }
          }
        ).reshapeAs(n)
      },
    }),
    al = An({
      reverse1d_: function (t) {
        var e = mn(t, 'x', 'reverse')
        return (
          C(1 === e.rank, function () {
            return (
              'Error in reverse1D: x must be rank 1 but got rank ' +
              e.rank +
              '.'
            )
          }),
          ol(e, 0)
        )
      },
    }),
    il = An({
      reverse2d_: function (t, e) {
        var n = mn(t, 'x', 'reverse')
        return (
          C(2 === n.rank, function () {
            return (
              'Error in reverse2D: x must be rank 2 but got rank ' +
              n.rank +
              '.'
            )
          }),
          ol(n, e)
        )
      },
    }),
    sl = An({
      reverse3d_: function (t, e) {
        var n = mn(t, 'x', 'reverse')
        return (
          C(3 === n.rank, function () {
            return (
              'Error in reverse3D: x must be rank 3 but got rank ' +
              n.rank +
              '.'
            )
          }),
          ol(n, e)
        )
      },
    }),
    ul = An({
      reverse4d_: function (t, e) {
        var n = mn(t, 'x', 'reverse')
        return (
          C(4 === n.rank, function () {
            return (
              'Error in reverse4D: x must be rank 4 but got rank ' +
              n.rank +
              '.'
            )
          }),
          ol(n, e)
        )
      },
    })
  function cl(t, e, n, r, o, a) {
    var i = mn(t, 'x', 'maxPool'),
      s = i,
      u = !1
    3 === i.rank &&
      ((u = !0), (s = i.as4D(1, i.shape[0], i.shape[1], i.shape[2]))),
      null == r && (r = [1, 1]),
      C(4 === s.rank, function () {
        return (
          'Error in maxPool: input must be rank 4 but got rank ' + s.rank + '.'
        )
      }),
      C(Mo(n, r), function () {
        return (
          'Error in maxPool: Either strides or dilations must be 1. Got strides ' +
          n +
          " and dilations '" +
          r +
          "'"
        )
      }),
      null != a &&
        C(A(o), function () {
          return (
            'Error in maxPool: pad must be an integer when using, dimRoundingMode ' +
            a +
            ' but got pad ' +
            o +
            '.'
          )
        })
    var c = Io(s.shape, e, n, r, o, a)
    if (1 === c.filterWidth && 1 === c.filterHeight && S(c.inShape, c.outShape))
      return i.clone()
    var l = [s],
      h = Lt.runKernelFunc(
        function (t, e) {
          var n = t.maxPool(s, c)
          return e([s, n]), n
        },
        { x: s },
        function (t, a) {
          var i = a[0],
            s = a[1]
          return {
            x: function () {
              return (function (t, e, n, r, o, a, i, s) {
                var u = mn(t, 'dy', 'maxPoolBackprop'),
                  c = mn(e, 'input', 'maxPoolBackprop'),
                  l = mn(n, 'output', 'maxPoolBackprop')
                C(c.rank === u.rank, function () {
                  return (
                    'Rank of input (' +
                    c.rank +
                    ') does not match rank of dy (' +
                    u.rank +
                    ')'
                  )
                }),
                  null == a && (a = [1, 1])
                C(Mo(o, a), function () {
                  return (
                    'Error in maxPoolBackProp: Either strides or dilations must be 1. Got strides ' +
                    o +
                    " and dilations '" +
                    a +
                    "'"
                  )
                }),
                  C(4 === u.rank, function () {
                    return (
                      'Error in maxPoolBackprop: dy must be rank 4 but got rank ' +
                      u.rank +
                      '.'
                    )
                  }),
                  C(4 === c.rank, function () {
                    return (
                      'Error in maxPoolBackprop: input must be rank 4 but got rank ' +
                      c.rank +
                      '.'
                    )
                  }),
                  null != s &&
                    C(A(i), function () {
                      return (
                        'Error in maxPoolBackprop: pad must be an integer when using, dimRoundingMode ' +
                        s +
                        ' but got pad ' +
                        i +
                        '.'
                      )
                    })
                var h = Io(c.shape, r, o, a, i, s)
                return Lt.runKernelFunc(
                  function (t) {
                    return t.maxPoolBackprop(u, c, l, h)
                  },
                  { $dy: u, $input: c }
                )
              })(t, i, s, e, n, r, o)
            },
          }
        },
        'MaxPool',
        c,
        l
      )
    return u ? h.as3D(h.shape[1], h.shape[2], h.shape[3]) : h
  }
  function ll(t, e, n, r, o, a) {
    var i = mn(t, 'x', 'avgPool', 'float32')
    null == r && (r = [1, 1]),
      C(Mo(n, r), function () {
        return (
          'Error in avgPool: Either strides or dilations must be 1. Got strides ' +
          n +
          " and dilations '" +
          r +
          "'"
        )
      })
    var s = i,
      u = !1
    3 === i.rank &&
      ((u = !0), (s = i.as4D(1, i.shape[0], i.shape[1], i.shape[2]))),
      C(4 === s.rank, function () {
        return 'Error in avgPool: x must be rank 4 but got rank ' + s.rank + '.'
      }),
      null != a &&
        C(A(o), function () {
          return (
            'Error in avgPool: pad must be an integer when using, dimRoundingMode ' +
            a +
            ' but got pad ' +
            o +
            '.'
          )
        })
    var c = Io(s.shape, e, n, r, o, a)
    if (1 === c.filterWidth && 1 === c.filterHeight && S(c.inShape, c.outShape))
      return i.clone()
    var l = Lt.runKernelFunc(
      function (t) {
        return t.avgPool(s, c)
      },
      { x: s },
      function (t) {
        return {
          x: function () {
            return (function (t, e, n, r, o, a) {
              var i = mn(t, 'dy', 'avgPoolBackprop'),
                s = mn(e, 'input', 'avgPoolBackprop')
              C(s.rank === i.rank, function () {
                return (
                  'Rank of input (' +
                  s.rank +
                  ') does not match rank of dy (' +
                  i.rank +
                  ')'
                )
              }),
                null == o && (o = [1, 1])
              C(Mo(r, o), function () {
                return (
                  'Error in avgPoolBackprop: Either strides or dilations must be 1. Got strides ' +
                  r +
                  " and dilations '" +
                  o +
                  "'"
                )
              })
              var u = s,
                c = i,
                l = !1
              3 === s.rank &&
                ((l = !0),
                (u = s.as4D(1, s.shape[0], s.shape[1], s.shape[2])),
                (c = i.as4D(1, i.shape[0], i.shape[1], i.shape[2])))
              C(4 === c.rank, function () {
                return (
                  'Error in avgPoolBackprop: dy must be rank 4 but got rank ' +
                  c.rank +
                  '.'
                )
              }),
                C(4 === u.rank, function () {
                  return (
                    'Error in avgPoolBackprop: input must be rank 4 but got rank ' +
                    u.rank +
                    '.'
                  )
                })
              var h = Io(u.shape, n, r, o, a),
                f = Lt.runKernelFunc(
                  function (t) {
                    return t.avgPoolBackprop(c, u, h)
                  },
                  { dy4D: c, input4D: u }
                )
              if (l) return f.as3D(f.shape[1], f.shape[2], f.shape[3])
              return f
            })(t, s, e, n, r, o)
          },
        }
      },
      'AvgPool',
      c
    )
    return (
      (l = l.cast(i.dtype)), u ? l.as3D(l.shape[1], l.shape[2], l.shape[3]) : l
    )
  }
  var hl = An({
      maxPool_: function (t, e, n, r, o) {
        return cl(t, e, n, 1, r, o)
      },
    }),
    fl = An({
      avgPool_: function (t, e, n, r, o) {
        return ll(t, e, n, 1, r, o)
      },
    }),
    dl = An({
      pool_: function (t, e, n, r, o, a) {
        null == o && (o = [1, 1]),
          null == a && (a = 1),
          0 === r && (r = 'valid')
        var i = mn(t, 'x', 'maxPool'),
          s = i,
          u = !1
        3 === i.rank &&
          ((u = !0), (s = i.as4D(1, i.shape[0], i.shape[1], i.shape[2]))),
          C(Mo(a, o), function () {
            return (
              'Error in pool: Either strides or dilations must be 1. Got strides ' +
              a +
              " and dilations '" +
              o +
              "'"
            )
          })
        var c,
          l = Io(s.shape, e, a, o, r),
          h = [l.dilationHeight, l.dilationWidth]
        c =
          'same' === r
            ? (function (t, e) {
                var n = t
                    .map(function (t, n) {
                      return t + (t - 1) * (e[n] - 1)
                    })
                    .map(function (t) {
                      return t - 1
                    }),
                  r = n.map(function (t) {
                    return Math.floor(t / 2)
                  }),
                  o = n.map(function (t, e) {
                    return t - r[e]
                  })
                return n.map(function (t, e) {
                  return [r[e], o[e]]
                })
              })([l.filterHeight, l.filterWidth], h)
            : [
                [0, 0],
                [0, 0],
              ]
        var f = 1 === h[0] && 1 === h[1],
          d = (function (t, e, n) {
            var r = n.map(function (t) {
                return t[0]
              }),
              o = n.map(function (t) {
                return t[1]
              }),
              a = t.concat(r, o),
              i = e.map(function (t, e) {
                return (t - (a[e] % t)) % t
              }),
              s = o.map(function (t, e) {
                return t + i[e]
              }),
              u = e.map(function (t, e) {
                return [r[e], s[e]]
              }),
              c = e.map(function (t, e) {
                return [0, i[e]]
              })
            return [u, c]
          })([l.inHeight, l.inWidth], h, c),
          p = d[0],
          v = d[1],
          m = f ? r : 'valid',
          g = f ? s : Mr(s, h, p),
          y = (
            'avg' === n
              ? function () {
                  return ll(g, e, a, 1, m)
                }
              : function () {
                  return cl(g, e, a, 1, m)
                }
          )(),
          x = f ? y : vr(y, h, v)
        return u ? x.as3D(x.shape[1], x.shape[2], x.shape[3]) : x
      },
    }),
    pl = An({
      maxPool3d_: function (t, e, n, r, o, a, i) {
        void 0 === a && (a = 'NDHWC')
        var s = mn(t, 'x', 'maxPool3d'),
          u = s,
          c = !1
        4 === s.rank &&
          ((c = !0),
          (u = s.as5D(1, s.shape[0], s.shape[1], s.shape[2], s.shape[3]))),
          null == i && (i = [1, 1, 1]),
          C(5 === u.rank, function () {
            return (
              'Error in maxPool3d: x must be rank 5 but got rank ' +
              u.rank +
              '.'
            )
          }),
          C('NDHWC' === a, function () {
            return (
              'Error in maxPool3d: Only NDHWC is currently supported, but got dataFormat of ' +
              a
            )
          }),
          C(Mo(n, i), function () {
            return (
              'Error in maxPool3d: Either strides or dilations must be 1. Got strides ' +
              n +
              " and dilations '" +
              i +
              "'"
            )
          }),
          null != o &&
            C(A(r), function () {
              return (
                'Error in maxPool3d: pad must be an integer when using, dimRoundingMode ' +
                o +
                ' but got pad ' +
                r +
                '.'
              )
            })
        var l = ko(u.shape, e, n, i, r, o, a),
          h = Lt.runKernelFunc(
            function (t, e) {
              var n = t.maxPool3d(u, l)
              return e([u, n]), n
            },
            { x: u },
            function (t, a) {
              var s = a[0],
                u = a[1]
              return {
                x: function () {
                  return (function (t, e, n, r, o, a, i, s) {
                    var u = mn(t, 'dy', 'maxPool3dBackprop'),
                      c = mn(e, 'input', 'maxPool3dBackprop'),
                      l = mn(n, 'output', 'maxPool3dBackprop'),
                      h = u,
                      f = c,
                      d = l,
                      p = !1
                    4 === c.rank &&
                      ((p = !0),
                      (h = u.as5D(
                        1,
                        u.shape[0],
                        u.shape[1],
                        u.shape[2],
                        u.shape[3]
                      )),
                      (f = c.as5D(
                        1,
                        c.shape[0],
                        c.shape[1],
                        c.shape[2],
                        c.shape[3]
                      )),
                      (d = l.as5D(
                        1,
                        l.shape[0],
                        l.shape[1],
                        l.shape[2],
                        l.shape[3]
                      )))
                    C(5 === h.rank, function () {
                      return (
                        'Error in maxPool3dBackprop: dy must be rank 5 but got rank ' +
                        h.rank +
                        '.'
                      )
                    }),
                      C(5 === f.rank, function () {
                        return (
                          'Error in maxPool3dBackprop: input must be rank 5 but got rank ' +
                          f.rank +
                          '.'
                        )
                      }),
                      C(5 === d.rank, function () {
                        return (
                          'Error in maxPool3dBackprop: output must be rank 5 but got rank ' +
                          d.rank +
                          '.'
                        )
                      }),
                      null == a && (a = [1, 1, 1])
                    C(Mo(o, a), function () {
                      return (
                        'Error in maxPool3dBackprop: Either strides or dilations must be 1. Got strides ' +
                        o +
                        " and dilations '" +
                        a +
                        "'"
                      )
                    }),
                      null != s &&
                        C(A(i), function () {
                          return (
                            'Error in maxPool3dBackprop: pad must be an integer when using, dimRoundingMode ' +
                            s +
                            ' but got pad ' +
                            i +
                            '.'
                          )
                        })
                    var v = ko(f.shape, r, o, a, i, s),
                      m = Lt.runKernelFunc(
                        function (t) {
                          return t.maxPool3dBackprop(h, f, d, v)
                        },
                        { dy5D: h, input5D: f }
                      )
                    if (p)
                      return m.as4D(
                        m.shape[1],
                        m.shape[2],
                        m.shape[3],
                        m.shape[4]
                      )
                    return m
                  })(t, s, u, e, n, i, r, o)
                },
              }
            }
          )
        return c ? h.as4D(h.shape[1], h.shape[2], h.shape[3], h.shape[4]) : h
      },
    }),
    vl = An({
      avgPool3d_: function (t, e, n, r, o, a, i) {
        void 0 === a && (a = 'NDHWC')
        var s = mn(t, 'x', 'avgPool3d', 'float32'),
          u = s,
          c = !1
        4 === s.rank &&
          ((c = !0),
          (u = s.as5D(1, s.shape[0], s.shape[1], s.shape[2], s.shape[3]))),
          null == i && (i = [1, 1, 1]),
          C(5 === u.rank, function () {
            return (
              'Error in avgPool3d: x must be rank 5 but got rank ' +
              u.rank +
              '.'
            )
          }),
          C('NDHWC' === a, function () {
            return (
              'Error in avgPool3d: Only NDHWC is currently supported, but got dataFormat of ' +
              a
            )
          }),
          C(Mo(n, i), function () {
            return (
              'Error in avgPool3d: Either strides or dilations must be 1. Got strides ' +
              n +
              " and dilations '" +
              i +
              "'"
            )
          }),
          null != o &&
            C(A(r), function () {
              return (
                'Error in avgPool3d: pad must be an integer when using, dimRoundingMode ' +
                o +
                ' but got pad ' +
                r +
                '.'
              )
            })
        var l = ko(u.shape, e, n, i, r, o, a),
          h = Lt.runKernelFunc(
            function (t) {
              return t.avgPool3d(u, l)
            },
            { x: u },
            function (t) {
              return {
                x: function () {
                  return (function (t, e, n, r, o, a, i) {
                    var s = mn(t, 'dy', 'avgPool3dBackprop'),
                      u = mn(e, 'input', 'avgPool3dBackprop'),
                      c = s,
                      l = u,
                      h = !1
                    4 === u.rank &&
                      ((h = !0),
                      (c = s.as5D(
                        1,
                        s.shape[0],
                        s.shape[1],
                        s.shape[2],
                        s.shape[3]
                      )),
                      (l = u.as5D(
                        1,
                        u.shape[0],
                        u.shape[1],
                        u.shape[2],
                        u.shape[3]
                      )))
                    C(5 === c.rank, function () {
                      return (
                        'Error in avgPool3dBackprop: dy must be rank 5 but got rank ' +
                        c.rank +
                        '.'
                      )
                    }),
                      C(5 === l.rank, function () {
                        return (
                          'Error in avgPool3dBackprop: input must be rank 5 but got rank ' +
                          l.rank +
                          '.'
                        )
                      }),
                      null == o && (o = [1, 1, 1])
                    C(Mo(r, o), function () {
                      return (
                        'Error in avgPool3dBackprop: Either strides or dilations must be 1. Got strides ' +
                        r +
                        " and dilations '" +
                        o +
                        "'"
                      )
                    }),
                      null != i &&
                        C(A(a), function () {
                          return (
                            'Error in maxPool3dBackprop: pad must be an integer when using, dimRoundingMode ' +
                            i +
                            ' but got pad ' +
                            a +
                            '.'
                          )
                        })
                    var f = ko(l.shape, n, r, o, a, i),
                      d = Lt.runKernelFunc(
                        function (t) {
                          return t.avgPool3dBackprop(c, l, f)
                        },
                        { dy5D: c, input5D: l }
                      )
                    if (h)
                      return d.as4D(
                        d.shape[1],
                        d.shape[2],
                        d.shape[3],
                        d.shape[4]
                      )
                    return d
                  })(t, u, e, n, i, r, o)
                },
              }
            }
          )
        return (
          (h = h.cast(u.dtype)),
          c ? h.as4D(h.shape[1], h.shape[2], h.shape[3], h.shape[4]) : h
        )
      },
    })
  var ml = An({
      slice_: function (t, e, n) {
        var r,
          o,
          a = mn(t, 'x', 'slice')
        if (0 === a.rank) throw new Error('Slicing scalar is not possible')
        ;(r =
          'number' == typeof e
            ? [e].concat(new Array(a.rank - 1).fill(0))
            : e.length < a.rank
            ? e.concat(new Array(a.rank - e.length).fill(0))
            : e.slice()).forEach(function (t) {
          C(-1 !== t, function () {
            return 'slice() does not support negative begin indexing.'
          })
        }),
          (o = (o =
            null == n
              ? new Array(a.rank).fill(-1)
              : 'number' == typeof n
              ? [n].concat(new Array(a.rank - 1).fill(-1))
              : n.length < a.rank
              ? n.concat(new Array(a.rank - n.length).fill(-1))
              : n).map(function (t, e) {
            return t >= 0
              ? t
              : (C(-1 === t, function () {
                  return (
                    'Negative size values should be exactly -1 but got ' +
                    t +
                    ' for the slice() size at index ' +
                    e +
                    '.'
                  )
                }),
                a.shape[e] - r[e])
          })),
          eo(a, r, o)
        var i = a.shape,
          s = { begin: r, size: o }
        return Lt.runKernelFunc(
          function (t) {
            return t.slice(a, r, o)
          },
          { x: a },
          function (t) {
            for (var e = [], n = 0; n < t.rank; n++)
              e.push([r[n], i[n] - r[n] - o[n]])
            return {
              x: function () {
                return t.pad(e)
              },
            }
          },
          'Slice',
          s
        )
      },
    }),
    gl = An({
      slice1d_: function (t, e, n) {
        var r = mn(t, 'x', 'slice1d')
        return (
          C(1 === r.rank, function () {
            return (
              'slice1d expects a rank-1 tensor, but got a rank-' +
              r.rank +
              ' tensor'
            )
          }),
          ml(r, [e], [n])
        )
      },
    }),
    yl = An({
      slice2d_: function (t, e, n) {
        var r = mn(t, 'x', 'slice2d')
        return (
          C(2 === r.rank, function () {
            return (
              'slice2d expects a rank-2 tensor, but got a rank-' +
              r.rank +
              ' tensor'
            )
          }),
          ml(r, e, n)
        )
      },
    }),
    xl = An({
      slice3d_: function (t, e, n) {
        var r = mn(t, 'x', 'slice3d')
        return (
          C(3 === r.rank, function () {
            return (
              'slice3d expects a rank-3 tensor, but got a rank-' +
              r.rank +
              ' tensor'
            )
          }),
          ml(r, e, n)
        )
      },
    }),
    bl = An({
      slice4d_: function (t, e, n) {
        var r = mn(t, 'x', 'slice4d')
        return (
          C(4 === r.rank, function () {
            return (
              'slice4d expects a rank-4 tensor, but got a rank-' +
              r.rank +
              ' tensor'
            )
          }),
          ml(r, e, n)
        )
      },
    })
  function wl(t, e, n, r, o) {
    return (
      e.rank < n.rank && (e = e.reshape(wn(e.shape, r))),
      t.rank < n.rank && (t = t.reshape(wn(t.shape, r))),
      {
        x: function () {
          var r = t.mul(n.equal(e).cast(t.dtype))
          return null == o ? r : r.transpose(o)
        },
      }
    )
  }
  var Cl = An({
      all_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'all', 'bool'),
          o = O(e, r.shape),
          a = o,
          i = En(a, r.rank)
        null != i && ((r = r.transpose(i)), (a = In(a.length, r.rank)))
        var s = Lt.runKernelFunc(
          function (t) {
            return t.all(r, a)
          },
          { $x: r }
        )
        if (n) {
          var u = wn(s.shape, o)
          return s.reshape(u)
        }
        return s
      },
    }),
    El = An({
      any_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'any', 'bool'),
          o = O(e, r.shape),
          a = o,
          i = En(a, r.rank)
        null != i && ((r = r.transpose(i)), (a = In(a.length, r.rank)))
        var s = Lt.runKernelFunc(
          function (t) {
            return t.any(r, a)
          },
          { $x: r }
        )
        if (n) {
          var u = wn(s.shape, o)
          return s.reshape(u)
        }
        return s
      },
    }),
    Rl = An({
      argMax_: function (t, e) {
        void 0 === e && (e = 0)
        var n = mn(t, 'x', 'argMax')
        null == e && (e = 0)
        var r = O(e, n.shape),
          o = En(r, n.rank)
        null != o && ((n = n.transpose(o)), (r = In(r.length, n.rank)))
        var a = { axis: r[0] },
          i = [n]
        return Lt.runKernelFunc(
          function (t, e) {
            var o = t.argMax(n, r[0])
            return e([n]), o
          },
          { x: n },
          function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return Xn(n)
              },
            }
          },
          'ArgMax',
          a,
          i
        )
      },
    }),
    Il = An({
      argMin_: function (t, e) {
        void 0 === e && (e = 0)
        var n = mn(t, 'x', 'argMin')
        null == e && (e = 0)
        var r = O(e, n.shape),
          o = En(r, n.rank)
        return (
          null != o && ((n = n.transpose(o)), (r = In(r.length, n.rank))),
          Lt.runKernelFunc(
            function (t, e) {
              var o = t.argMin(n, r[0])
              return e([n]), o
            },
            { $x: n },
            function (t, e) {
              var n = e[0]
              return {
                $x: function () {
                  return Xn(n)
                },
              }
            }
          )
        )
      },
    }),
    kl = An({
      logSumExp_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'logSumExp'),
          o = O(e, r.shape),
          a = r.max(o, !0),
          i = r.sub(a).exp().sum(o).log(),
          s = a.reshape(i.shape).add(i)
        if (n) {
          var u = wn(s.shape, o)
          return s.reshape(u)
        }
        return s
      },
    }),
    Sl = An({
      max_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'max'),
          o = r,
          a = O(e, r.shape),
          i = a,
          s = En(i, r.rank)
        null != s && ((r = r.transpose(s)), (i = In(i.length, r.rank)))
        var u = [r],
          c = Lt.runKernelFunc(
            function (t, e) {
              var n = t.max(r, i)
              return e([o, n]), n
            },
            { x: r },
            function (t, e) {
              return wl(t, e[1], e[0], a, s)
            },
            'Max',
            { axes: i },
            u,
            [!0]
          )
        if (n) {
          var l = wn(c.shape, a)
          c = c.reshape(l)
        }
        return c
      },
    }),
    Al = An({
      mean_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'mean'),
          o = O(e, r.shape),
          a = k(bn(r.shape, o)[1])
        return vo(function (t) {
          var r = On(a)
          return {
            value: (r.dtype === t.dtype ? t : t.cast(r.dtype)).div(r).sum(e, n),
            gradFunc: function (e) {
              var n = t.shape.slice()
              return (
                o.forEach(function (t) {
                  n[t] = 1
                }),
                e.reshape(n).mul(zn(t.shape, 'float32')).div(a)
              )
            },
          }
        })(r)
      },
    }),
    Dl = An({
      min_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'min'),
          o = r,
          a = O(e, r.shape),
          i = a,
          s = En(i, r.rank)
        null != s && ((r = r.transpose(s)), (i = In(i.length, r.rank)))
        var u = [r],
          c = Lt.runKernelFunc(
            function (t, e) {
              var n = t.min(r, i)
              return e([o, n]), n
            },
            { x: r },
            function (t, e) {
              return wl(t, e[1], e[0], a, s)
            },
            'Min',
            { axes: i },
            u,
            [!0]
          )
        if (n) {
          var l = wn(c.shape, a)
          c = c.reshape(l)
        }
        return c
      },
    }),
    Tl = An({
      moments_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = O(e, (t = mn(t, 'x', 'moments')).shape),
          o = t.mean(r, n),
          a = o.shape
        n || (a = wn(o.shape, r))
        var i = t.toFloat().sub(o.reshape(a)).square()
        return { mean: o, variance: i.mean(r, n) }
      },
    }),
    Nl = An({
      sum_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'sum')
        'bool' === r.dtype && (r = r.toInt())
        var o = O(e, r.shape)
        return vo(function (t) {
          var e = En(o, t.rank),
            r = o,
            a = t
          null != e && ((a = t.transpose(e)), (r = In(r.length, t.rank)))
          var i = function (e) {
              var n = t.shape.slice()
              return (
                o.forEach(function (t) {
                  n[t] = 1
                }),
                e.reshape(n).mul(zn(t.shape, 'float32'))
              )
            },
            s = { axes: r },
            u = Lt.runKernelFunc(
              function (t) {
                return t.sum(a, r)
              },
              { x: a },
              function (t) {
                return {
                  x: function () {
                    return i(t)
                  },
                }
              },
              'Sum',
              s
            )
          if (n) {
            var c = wn(u.shape, o)
            u = u.reshape(c)
          }
          return { value: u, gradFunc: i }
        })(r)
      },
    }),
    Fl = An({
      prod_: function (t, e, n) {
        void 0 === e && (e = null), void 0 === n && (n = !1)
        var r = mn(t, 'x', 'prod')
        'bool' === r.dtype && (r = r.toInt())
        var o = O(e, r.shape),
          a = En(o, r.rank),
          i = o,
          s = r
        null != a && ((s = r.transpose(a)), (i = In(i.length, r.rank)))
        var u = Lt.runKernelFunc(
          function (t) {
            return t.prod(s, i)
          },
          { permutedX: s }
        )
        if (n) {
          var c = wn(u.shape, o)
          u = u.reshape(c)
        }
        return u
      },
    })
  var _l = An({
      elu_: function (t) {
        var e = mn(t, 'x', 'elu')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.elu(e)
            return n([r]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                return Lt.runKernelFunc(
                  function (e) {
                    return e.eluDer(t, n)
                  },
                  { dy: t, y: n }
                )
              },
            }
          }
        )
      },
    }),
    Ol = An({
      leakyRelu_: function (t, e) {
        void 0 === e && (e = 0.2)
        var n = mn(t, 'x', 'leakyRelu')
        return hc(On(e).mul(n), n)
      },
    }),
    Ml = An({
      prelu_: function (t, e) {
        var n = mn(t, 'x', 'prelu'),
          r = mn(e, 'alpha', 'prelu')
        return Lt.runKernelFunc(
          function (t, e) {
            var o = t.prelu(n, r)
            return e([n, r]), o
          },
          { x: n, alpha: r },
          function (t, e) {
            var n = e[0],
              r = e[1],
              o = n.greater(0)
            return {
              x: function () {
                return ec(o, t, t.mul(r))
              },
              alpha: function () {
                var e = ec(o, Xn(t), t.mul(n)),
                  a = Eo(r.shape, t.shape)
                return a.length > 0 && (e = e.sum(a)), e.reshape(r.shape)
              },
            }
          },
          'Prelu'
        )
      },
    }),
    Bl = An({
      relu_: function (t) {
        var e = mn(t, 'x', 'relu')
        return 'bool' === e.dtype
          ? e.toInt()
          : Lt.runKernelFunc(
              function (t, n) {
                var r = t.relu(e)
                return n([e]), r
              },
              { x: e },
              function (t, e) {
                var n = e[0]
                return {
                  x: function () {
                    return t.mulStrict(n.step().toFloat())
                  },
                }
              },
              'Relu'
            )
      },
    }),
    Pl = An({
      relu6_: function (t) {
        var e = mn(t, 'x', 'relu6')
        return 'bool' === e.dtype
          ? e.toInt()
          : Lt.runKernelFunc(
              function (t, n) {
                var r = t.relu6(e)
                return n([e]), r
              },
              { x: e },
              function (t, e) {
                var n = e[0],
                  r = n.lessEqual(6).mul(n.step())
                return {
                  x: function () {
                    return t.mulStrict(r.toFloat())
                  },
                }
              },
              'Relu6'
            )
      },
    }),
    Ll = An({
      selu_: function (t) {
        var e = mn(t, 'x', 'selu')
        return Lt.runKernelFunc(
          function (t, n) {
            var r = t.selu(e)
            return n([e]), r
          },
          { $x: e },
          function (t, e) {
            var n = e[0]
            return {
              $x: function () {
                var e = n.greater(On(0)),
                  r = On(bs),
                  o = On(ws),
                  a = t.mul(o),
                  i = t.mul(r).mul(n.toFloat().exp())
                return ec(e, a, i)
              },
            }
          }
        )
      },
    })
  var Wl = An({
    transpose_: function (t, e) {
      var n = mn(t, 'x', 'transpose')
      if (
        (null == e &&
          (e = n.shape
            .map(function (t, e) {
              return e
            })
            .reverse()),
        C(n.rank === e.length, function () {
          return (
            'Error in transpose: rank of input ' +
            n.rank +
            ' must match length of perm ' +
            e +
            '.'
          )
        }),
        e.forEach(function (t) {
          C(t >= 0 && t < n.rank, function () {
            return (
              "All entries in 'perm' must be between 0 and " +
              (n.rank - 1) +
              ' but got ' +
              e
            )
          })
        }),
        n.rank <= 1)
      )
        return n.clone()
      var r = { perm: e }
      return Lt.runKernelFunc(
        function (t) {
          return t.transpose(n, e)
        },
        { x: n },
        function (t) {
          var n = Rn(e)
          return {
            x: function () {
              return t.transpose(n)
            },
          }
        },
        'Transpose',
        r
      )
    },
  })
  var Ul = An({
    localResponseNormalization_: function (t, e, n, r, o) {
      void 0 === e && (e = 5),
        void 0 === n && (n = 1),
        void 0 === r && (r = 1),
        void 0 === o && (o = 0.5)
      var a = mn(t, 'x', 'localResponseNormalization')
      C(4 === a.rank || 3 === a.rank, function () {
        return (
          'Error in localResponseNormalization: x must be rank 3 or 4 but got\n               rank ' +
          a.rank +
          '.'
        )
      }),
        C(A(e), function () {
          return (
            'Error in localResponseNormalization: depthRadius must be an integer but got depthRadius ' +
            e +
            '.'
          )
        })
      var i = a,
        s = !1
      3 === a.rank &&
        ((s = !0), (i = a.as4D(1, a.shape[0], a.shape[1], a.shape[2])))
      var u = Lt.runKernelFunc(
        function (t, a) {
          var s = t.localResponseNormalization4D(i, e, n, r, o)
          return a([i, s]), s
        },
        { x4D: i },
        function (t, a) {
          var i = a[0],
            s = a[1]
          return {
            x4D: function () {
              return Lt.runKernelFunc(function (a) {
                return a.LRNGrad(t, i, s, e, n, r, o)
              }, {})
            },
          }
        }
      )
      return s ? u.as3D(u.shape[1], u.shape[2], u.shape[3]) : u
    },
  })
  var Vl = An({
    norm_: function (t, e, n, r) {
      void 0 === e && (e = 'euclidean'),
        void 0 === n && (n = null),
        void 0 === r && (r = !1)
      var o = (function t(e, n, r) {
          void 0 === r && (r = null)
          if (0 === e.rank) return e.abs()
          if (1 !== e.rank && null === r) return t(e.reshape([-1]), n, r)
          if (
            1 === e.rank ||
            'number' == typeof r ||
            (Array.isArray(r) && 1 === r.length)
          ) {
            if (1 === n) return e.abs().sum(r)
            if (n === 1 / 0) return e.abs().max(r)
            if (n === -1 / 0) return e.abs().min(r)
            if ('euclidean' === n || 2 === n)
              return e.abs().pow(On(2, 'int32')).sum(r).sqrt()
            throw new Error('Error in norm: invalid ord value: ' + n)
          }
          if (Array.isArray(r) && 2 === r.length) {
            if (1 === n)
              return e
                .abs()
                .sum(r[0])
                .max(r[1] - 1)
            if (n === 1 / 0) return e.abs().sum(r[1]).max(r[0])
            if (n === -1 / 0) return e.abs().sum(r[1]).min(r[0])
            if ('fro' === n || 'euclidean' === n)
              return e.square().sum(r).sqrt()
            throw new Error('Error in norm: invalid ord value: ' + n)
          }
          throw new Error('Error in norm: invalid axis: ' + r)
        })((t = mn(t, 'x', 'norm')), e, n),
        a = o.shape
      if (r) {
        var i = O(n, t.shape)
        a = wn(o.shape, i)
      }
      return o.reshape(a)
    },
  })
  var zl = An({
      basicLSTMCell_: function (t, e, n, r, o, a) {
        var i = mn(t, 'forgetBias', 'basicLSTMCell'),
          s = mn(e, 'lstmKernel', 'basicLSTMCell'),
          u = mn(n, 'lstmBias', 'basicLSTMCell'),
          c = mn(r, 'data', 'basicLSTMCell'),
          l = mn(o, 'c', 'basicLSTMCell'),
          h = mn(a, 'h', 'basicLSTMCell'),
          f = c.concat(h, 1).matMul(s).add(u),
          d = f.shape[0],
          p = f.shape[1] / 4,
          v = [d, p],
          m = f.slice([0, 0], v),
          g = f.slice([0, p], v),
          y = f.slice([0, 2 * p], v),
          x = f.slice([0, 3 * p], v),
          b = m
            .sigmoid()
            .mulStrict(g.tanh())
            .addStrict(l.mulStrict(i.add(y).sigmoid())),
          w = b.tanh().mulStrict(x.sigmoid())
        return [b, w]
      },
    }),
    Gl = An({
      multiRNNCell_: function (t, e, n, r) {
        for (
          var o = mn(e, 'data', 'multiRNNCell'),
            a = gn(n, 'c', 'multiRNNCell'),
            i = gn(r, 'h', 'multiRNNCell'),
            s = o,
            u = [],
            c = 0;
          c < t.length;
          c++
        ) {
          var l = t[c](s, a[c], i[c])
          u.push(l[0]), u.push(l[1]), (s = l[1])
        }
        var h = [],
          f = []
        for (c = 0; c < u.length; c += 2) h.push(u[c]), f.push(u[c + 1])
        return [h, f]
      },
    })
  var Hl = An({
    movingAverage_: function (t, e, n, r, o) {
      void 0 === o && (o = !0)
      var a = mn(t, 'v', 'movingAverage'),
        i = mn(e, 'x', 'movingAverage'),
        s = mn(n, 'decay', 'movingAverage')
      Ft(a, i),
        C(S(a.shape, i.shape), function () {
          return 'Shape mismatch in v and x'
        })
      var u = On(1),
        c = u.sub(s),
        l = i.sub(a).mul(c)
      if (o) {
        C(null != r, function () {
          return 'When using zeroDebias: true, step is required.'
        })
        var h = mn(r, 'step', 'movingAverage')
        l = l.div(u.sub(xc(s, h)))
      }
      return a.add(l)
    },
  })
  var ql = An({
    stridedSlice_: function (t, e, n, r, o, a, i, s, u) {
      if (
        (void 0 === o && (o = 0),
        void 0 === a && (a = 0),
        void 0 === i && (i = 0),
        void 0 === s && (s = 0),
        void 0 === u && (u = 0),
        null == r && (r = new Array(e.length)),
        0 !== i)
      )
        throw new Error('ellipsis mask is not yet supported')
      var c = mn(t, 'x', 'stridedSlice'),
        l = no(s),
        h = c.shape.slice()
      l.forEach(function (t) {
        ;(e[t] = 0), (n[t] = 1), h.splice(t, 0, 1)
      }),
        (c = c.reshape(h))
      for (var f = 0; f < c.rank; f++)
        (e[f] = oo(o, e, r, c.shape, f)),
          (n[f] = ao(a, n, r, c.shape, f)),
          (r[f] = r[f] || 1)
      var d = no(u)
      d.forEach(function (t) {
        ;(n[t] = e[t] + 1), (r[t] = 1)
      })
      var p = ro(e, n, r),
        v = p.filter(function (t, e) {
          return -1 === d.indexOf(e)
        })
      return r.every(function (t) {
        return 1 === t
      })
        ? ml(c, e, p).reshape(v)
        : Lt.runKernelFunc(
            function (t) {
              return t.stridedSlice(c, e, n, r)
            },
            { $x: c }
          ).reshape(v)
    },
  })
  var Kl = An({
    topk_: function (t, e, n) {
      void 0 === e && (e = 1), void 0 === n && (n = !0)
      var r = mn(t, 'x', 'topk')
      if (0 === r.rank)
        throw new Error('topk() expects the input to be of rank 1 or higher')
      var o = r.shape[r.shape.length - 1]
      if (e > o)
        throw new Error(
          "'k' passed to topk() must be <= the last dimension (" +
            o +
            ') but got ' +
            e
        )
      var a = Lt.runKernelFunc(
        function (t) {
          return t.topk(r, e, n)
        },
        { $x: r }
      )
      return { values: a[0], indices: a[1] }
    },
  })
  var jl = An({
    scatterND_: function (t, e, n) {
      var r = mn(t, 'indices', 'scatterND', 'int32'),
        o = mn(e, 'updates', 'scatterND')
      return (
        Jr(o, r, n),
        Lt.runKernelFunc(
          function (t) {
            return t.scatterND(r, o, n)
          },
          { indices: r, updates: o },
          null,
          'ScatterNd',
          { shape: n }
        )
      )
    },
  })
  var Xl = An({
      fft_: function (t) {
        C('complex64' === t.dtype, function () {
          return (
            'The dtype for tf.spectral.fft() must be complex64 but got ' +
            t.dtype +
            '.'
          )
        })
        var e = t.shape[t.shape.length - 1],
          n = t.size / e,
          r = t.as2D(n, e)
        return Lt.runKernelFunc(
          function (t) {
            return t.fft(r)
          },
          { input: t }
        ).reshape(t.shape)
      },
    }),
    Yl = An({
      ifft_: function (t) {
        C('complex64' === t.dtype, function () {
          return (
            'The dtype for tf.spectral.ifft() must be complex64 but got ' +
            t.dtype +
            '.'
          )
        })
        var e = t.shape[t.shape.length - 1],
          n = t.size / e,
          r = t.as2D(n, e)
        return Lt.runKernelFunc(
          function (t) {
            return t.ifft(r)
          },
          { input: t }
        ).reshape(t.shape)
      },
    }),
    $l = An({
      rfft_: function (t, e) {
        C('float32' === t.dtype, function () {
          return 'The dtype for rfft() must be real value but got ' + t.dtype
        })
        var n,
          r = t.shape[t.shape.length - 1],
          o = t.size / r
        if (null != e && e < r) {
          var a = t.shape.map(function (t) {
              return 0
            }),
            i = t.shape.map(function (t) {
              return t
            })
          ;(i[t.shape.length - 1] = e), (n = t.slice(a, i)), (r = e)
        } else if (null != e && e > r) {
          var s = t.shape.map(function (t) {
            return t
          })
          ;(s[t.shape.length - 1] = e - r),
            (n = t.concat(Gn(s), t.shape.length - 1)),
            (r = e)
        } else n = t
        var u = n.zerosLike(),
          c = Dn(n, u).as2D(o, r),
          l = Xl(c),
          h = Math.floor(r / 2) + 1,
          f = Tn(l),
          d = Nn(l),
          p = f.split([h, r - h], f.shape.length - 1),
          v = d.split([h, r - h], d.shape.length - 1),
          m = n.shape.slice()
        return (m[n.shape.length - 1] = h), Dn(p[0], v[0]).reshape(m)
      },
    }),
    Ql = An({
      irfft_: function (t) {
        var e = t.shape[t.shape.length - 1],
          n = t.size / e
        if (e <= 2) {
          var r = t.as2D(n, e),
            o = Yl(r)
          return Tn(o)
        }
        var a = [n, 2 * (e - 1)],
          i = Tn(t).as2D(n, e),
          s = Nn(t).as2D(n, e),
          u = i.slice([0, 1], [n, e - 2]).reverse(1),
          c = s
            .slice([0, 1], [n, e - 2])
            .reverse(1)
            .mul(On(-1)),
          l = i.concat(u, 1),
          h = s.concat(c, 1)
        return (r = Dn(l, h).as2D(a[0], a[1])), (o = Yl(r)), Tn(o)
      },
    }),
    Jl = Object.freeze({ fft: Xl, ifft: Yl, rfft: $l, irfft: Ql })
  var Zl = An({
    sparseToDense_: function (t, e, n, r) {
      void 0 === r && (r = 0)
      var o = mn(t, 'sparseIndices', 'sparseToDense', 'int32'),
        a = mn(e, 'sparseValues', 'sparseToDense'),
        i = mn(r, 'defaultValue', 'sparseToDense', a.dtype)
      return (
        (function (t, e, n, r) {
          if ('int32' !== t.dtype)
            throw new Error(
              'tf.sparseToDense() expects the indices to be int32 type, but the dtype was ' +
                t.dtype +
                '.'
            )
          if (t.rank > 2)
            throw new Error(
              'sparseIndices should be a scalar, vector, or matrix, but got shape ' +
                t.shape +
                '.'
            )
          var o = t.rank > 0 ? t.shape[0] : 1,
            a = t.rank > 1 ? t.shape[1] : 1
          if (n.length !== a)
            throw new Error(
              'outputShape has incorrect number of elements:, ' +
                n.length +
                ', should be: ' +
                a +
                '.'
            )
          var i = e.size
          if (0 !== e.rank && (1 !== e.rank || i !== o))
            throw new Error(
              'sparseValues has incorrect shape ' +
                e.shape +
                ', should be [] or [' +
                o +
                ']'
            )
          if (e.dtype !== r.dtype)
            throw new Error('sparseValues.dtype must match defaultValues.dtype')
        })(o, a, n, i),
        Lt.runKernelFunc(
          function (t) {
            return t.sparseToDense(o, a, n, i)
          },
          { $sparseIndices: o, $sparseValues: a, $defaultValue: i }
        )
      )
    },
  })
  var th = An({
    gatherND_: function (t, e) {
      var n = mn(e, 'indices', 'gatherND', 'int32'),
        r = mn(t, 'x', 'gatherND')
      return Lt.runKernelFunc(
        function (t) {
          return t.gatherND(r, n)
        },
        { x: r, indices: n },
        null,
        'GatherNd'
      )
    },
  })
  var eh = An({
    diag_: function (t) {
      var e = mn(t, 'x', 'diag').flatten(),
        n = t.shape.concat(t.shape)
      return Lt.runKernelFunc(
        function (t) {
          return t.diag(e)
        },
        { $x: e }
      ).reshape(n)
    },
  })
  var nh = An({
    dropout_: function (t, e, n, r) {
      var o = mn(t, 'x', 'dropout')
      if (
        (C('float32' === o.dtype, function () {
          return (
            "x has to be a floating point tensor since it's going to be scaled, but got a " +
            o.dtype +
            ' tensor instead.'
          )
        }),
        C(e >= 0 && e < 1, function () {
          return 'rate must be a float in the range [0, 1), but got ' + e + '.'
        }),
        0 === e)
      )
        return t instanceof wt ? o.clone() : o
      var a = (function (t, e) {
          if (null == e) return t.shape.slice()
          if (S(t.shape, e)) return e
          if (t.shape.length === e.length) {
            for (var n = [], r = 0; r < t.shape.length; r++)
              null == e[r] && null != t.shape[r]
                ? n.push(t.shape[r])
                : n.push(e[r])
            return n
          }
          return e
        })(o, n),
        i = 1 - e,
        s = _r(a, 0, 1, 'float32', r).add(i).floor().div(i)
      return o.mul(s)
    },
  })
  function rh(t, e, n) {
    for (var r = 1 - (t % 2), o = new Float32Array(t), a = 0; a < t; ++a) {
      var i = (2 * Math.PI * a) / (t + r - 1)
      o[a] = e - n * Math.cos(i)
    }
    return Mn(o, 'float32')
  }
  var oh = An({
      hannWindow_: function (t) {
        return rh(t, 0.5, 0.5)
      },
    }),
    ah = An({
      hammingWindow_: function (t) {
        return rh(t, 0.54, 0.46)
      },
    }),
    ih = An({
      frame_: function (t, e, n, r, o) {
        void 0 === r && (r = !1), void 0 === o && (o = 0)
        for (var a = 0, i = []; a + e <= t.size; ) i.push(ml(t, a, e)), (a += n)
        if (r)
          for (; a < t.size; ) {
            var s = a + e - t.size,
              u = Yn([ml(t, a, e - s), Hn([s], o)])
            i.push(u), (a += n)
          }
        return 0 === i.length ? Bn([], [0, e]) : Yn(i).as2D(i.length, e)
      },
    }),
    sh = An({
      stft_: function (t, e, n, r, o) {
        var a
        void 0 === o && (o = oh),
          null == r &&
            ((a = e),
            (r = Math.floor(Math.pow(2, Math.ceil(Math.log(a) / Math.log(2))))))
        for (
          var i = ih(t, e, n), s = gc(i, o(e)), u = [], c = 0;
          c < i.shape[0];
          c++
        )
          u.push($l(s.slice([c, 0], [1, e]), r))
        return Yn(u)
      },
    }),
    uh = Object.freeze({
      hannWindow: oh,
      hammingWindow: ah,
      frame: ih,
      stft: sh,
    })
  var ch,
    lh = function (t, e, o) {
      return (
        void 0 === o && (o = 1),
        n(this, void 0, void 0, function () {
          var n, a, i, s, u, c, l, h, f, d, p, v, m, g
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                return (
                  (n = mn(t, 'predictions', 'inTopK')),
                  (a = mn(e, 'targets', 'inTopK')),
                  C(n.rank > 1, function () {
                    return (
                      'inTopK() expects the predictions to be of rank 2 or higher, but got ' +
                      n.rank
                    )
                  }),
                  C(n.rank - 1 === a.rank, function () {
                    return (
                      'predictions rank should be 1 larger than targets rank, but got predictions rank ' +
                      n.rank +
                      ' and targets rank ' +
                      a.rank
                    )
                  }),
                  E(
                    n.shape.slice(0, n.shape.length - 1),
                    a.shape,
                    "predictions's shape should be align with the targets' shape, except the last dimension."
                  ),
                  (i = n.shape[n.shape.length - 1]),
                  C(o > 0 && o <= i, function () {
                    return (
                      "'k' passed to inTopK() must be > 0 && <= the predictions last dimension (" +
                      i +
                      '), but got ' +
                      o
                    )
                  }),
                  [4, n.data()]
                )
              case 1:
                return (s = r.sent()), [4, a.data()]
              case 2:
                for (
                  u = r.sent(),
                    c = [s.length / i, i],
                    h = c[1],
                    f = B('bool', (l = c[0])),
                    d = 0;
                  d < l;
                  d++
                ) {
                  for (
                    p = d * h, v = s.subarray(p, p + h), m = [], g = 0;
                    g < v.length;
                    g++
                  )
                    m.push({ value: v[g], index: g })
                  for (
                    m.sort(function (t, e) {
                      return e.value - t.value
                    }),
                      f[d] = 0,
                      g = 0;
                    g < o;
                    g++
                  )
                    if (m[g].index === u[d]) {
                      f[d] = 1
                      break
                    }
                }
                return (
                  t !== n && n.dispose(),
                  e !== a && a.dispose(),
                  [2, Fn(f, a.shape, 'bool')]
                )
            }
          })
        })
      )
    }
  !(function (t) {
    ;(t[(t.NONE = 0)] = 'NONE'),
      (t[(t.MEAN = 1)] = 'MEAN'),
      (t[(t.SUM = 2)] = 'SUM'),
      (t[(t.SUM_BY_NONZERO_WEIGHTS = 3)] = 'SUM_BY_NONZERO_WEIGHTS')
  })(ch || (ch = {}))
  var hh = An({
      absoluteDifference_: function (t, e, n, r) {
        void 0 === r && (r = ch.SUM_BY_NONZERO_WEIGHTS)
        var o = mn(t, 'labels', 'absoluteDifference'),
          a = mn(e, 'predictions', 'absoluteDifference'),
          i = null
        null != n && (i = mn(n, 'weights', 'absoluteDifference')),
          E(o.shape, a.shape, 'Error in absoluteDifference: ')
        var s = o.sub(a).abs()
        return fh(s, i, r)
      },
    }),
    fh = An({
      computeWeightedLoss_: function (t, e, n) {
        void 0 === n && (n = ch.SUM_BY_NONZERO_WEIGHTS)
        var r = mn(t, 'losses', 'computeWeightedLoss'),
          o = null
        null != e && (o = mn(e, 'weights', 'computeWeightedLoss'))
        var a = null == o ? r : r.mul(o)
        if (n === ch.NONE) return a
        if (n === ch.SUM) return a.sum()
        if (n === ch.MEAN) {
          if (null == o) return a.mean()
          var i = r.size / o.size,
            s = a.sum().div(o.sum())
          return i > 1 ? s.div(On(i)) : s
        }
        if (n === ch.SUM_BY_NONZERO_WEIGHTS) {
          if (null == o) return a.sum().div(On(r.size))
          var u = o.mul(zn(r.shape)).notEqual(On(0)).sum().toFloat()
          return a.sum().div(u)
        }
        throw Error('Unknown reduction: ' + n)
      },
    }),
    dh = An({
      cosineDistance_: function (t, e, n, r, o) {
        void 0 === o && (o = ch.SUM_BY_NONZERO_WEIGHTS)
        var a = mn(t, 'labels', 'cosineDistance'),
          i = mn(e, 'predictions', 'cosineDistance'),
          s = null
        null != r && (s = mn(r, 'weights', 'cosineDistance')),
          E(a.shape, i.shape, 'Error in cosineDistance: ')
        var u = On(1).sub(a.mul(i).sum(n, !0))
        return fh(u, s, o)
      },
    }),
    ph = An({
      hingeLoss_: function (t, e, n, r) {
        void 0 === r && (r = ch.SUM_BY_NONZERO_WEIGHTS)
        var o = mn(t, 'labels', 'hingeLoss'),
          a = mn(e, 'predictions', 'hingeLoss'),
          i = null
        null != n && (i = mn(n, 'weights', 'hingeLoss')),
          E(o.shape, a.shape, 'Error in hingeLoss: ')
        var s = On(1)
        o = On(2).mul(o).sub(s)
        var u = s.sub(o.mul(a)).relu()
        return fh(u, i, r)
      },
    }),
    vh = An({
      huberLoss_: function (t, e, n, r, o) {
        void 0 === r && (r = 1), void 0 === o && (o = ch.SUM_BY_NONZERO_WEIGHTS)
        var a = mn(t, 'labels', 'huberLoss'),
          i = mn(e, 'predictions', 'huberLoss'),
          s = null
        null != n && (s = mn(n, 'weights', 'huberLoss')),
          E(a.shape, i.shape, 'Error in huberLoss: ')
        var u = On(r),
          c = i.sub(a).abs(),
          l = dc(c, u),
          h = c.sub(l),
          f = On(0.5).mul(l.square()).add(u.mul(h))
        return fh(f, s, o)
      },
    }),
    mh = An({
      logLoss_: function (t, e, n, r, o) {
        void 0 === r && (r = 1e-7),
          void 0 === o && (o = ch.SUM_BY_NONZERO_WEIGHTS)
        var a = mn(t, 'labels', 'logLoss'),
          i = mn(e, 'predictions', 'logLoss'),
          s = null
        null != n && (s = mn(n, 'weights', 'logLoss')),
          E(a.shape, i.shape, 'Error in logLoss: ')
        var u = On(1),
          c = On(r),
          l = a
            .mul(i.add(c).log())
            .neg()
            .sub(u.sub(a).mul(u.sub(i).add(c).log()))
        return fh(l, s, o)
      },
    }),
    gh = An({
      meanSquaredError_: function (t, e, n, r) {
        void 0 === r && (r = ch.SUM_BY_NONZERO_WEIGHTS)
        var o = mn(t, 'labels', 'meanSquaredError'),
          a = mn(e, 'predictions', 'meanSquaredError'),
          i = null
        null != n && (i = mn(n, 'weights', 'meanSquaredError')),
          E(o.shape, a.shape, 'Error in meanSquaredError: ')
        var s = o.squaredDifference(a)
        return fh(s, i, r)
      },
    }),
    yh = An({
      sigmoidCrossEntropy_: function (t, e, n, r, o) {
        void 0 === r && (r = 0), void 0 === o && (o = ch.SUM_BY_NONZERO_WEIGHTS)
        var a = mn(t, 'multiClassLabels', 'sigmoidCrossEntropy'),
          i = mn(e, 'logits', 'sigmoidCrossEntropy'),
          s = null
        if (
          (null != n && (s = mn(n, 'weights', 'sigmoidCrossEntropy')),
          E(a.shape, i.shape, 'Error in sigmoidCrossEntropy: '),
          r > 0)
        ) {
          var u = On(r),
            c = On(1),
            l = On(0.5)
          a = a.mul(c.sub(u)).add(l.mul(u))
        }
        var h = (function (t, e) {
          var n = mn(t, 'labels', 'sigmoidCrossEntropyWithLogits'),
            r = mn(e, 'logits', 'sigmoidCrossEntropyWithLogits')
          E(n.shape, r.shape, 'Error in sigmoidCrossEntropyWithLogits: ')
          var o = r.relu(),
            a = r.mul(n),
            i = r.abs().neg().exp().log1p()
          return o.sub(a).add(i)
        })(a, i)
        return fh(h, s, o)
      },
    }),
    xh = An({
      softmaxCrossEntropy_: function (t, e, n, r, o) {
        void 0 === r && (r = 0), void 0 === o && (o = ch.SUM_BY_NONZERO_WEIGHTS)
        var a = mn(t, 'onehotLabels', 'softmaxCrossEntropy'),
          i = mn(e, 'logits', 'softmaxCrossEntropy'),
          s = null
        if (
          (null != n && (s = mn(n, 'weights', 'softmaxCrossEntropy')),
          E(a.shape, i.shape, 'Error in softmaxCrossEntropy: '),
          r > 0)
        ) {
          var u = On(r),
            c = On(1),
            l = On(a.shape[1])
          a = a.mul(c.sub(u)).add(u.div(l))
        }
        var h = (function (t, e, n) {
          if (
            (void 0 === n && (n = -1),
            -1 === n && (n = e.rank - 1),
            n !== e.rank - 1)
          )
            throw Error(
              'Softmax cross entropy along a non-last dimension is not yet supported. Labels / logits was rank ' +
                e.rank +
                ' and dim was ' +
                n
            )
          return vo(function (t, e, r) {
            var o = e.logSumExp([n], !0),
              a = e.toFloat().sub(o)
            r([t, a])
            return {
              value: a.mul(t).neg().sum([n]),
              gradFunc: function (t, e) {
                var r = e[0],
                  o = e[1],
                  a = wn(t.shape, [n])
                return [
                  t.reshape(a).mul(r.toFloat().sub(o.exp())),
                  t.reshape(a).mul(o.exp().sub(r.toFloat())),
                ]
              },
            }
          })(t, e)
        })(a, i)
        return fh(h, s, o)
      },
    }),
    bh = Object.freeze({
      get Reduction() {
        return ch
      },
      absoluteDifference: hh,
      computeWeightedLoss: fh,
      cosineDistance: dh,
      hingeLoss: ph,
      huberLoss: vh,
      logLoss: mh,
      meanSquaredError: gh,
      sigmoidCrossEntropy: yh,
      softmaxCrossEntropy: xh,
    })
  function wh(t, e) {
    return (
      void 0 === e && (e = !1),
      Lt.tidy(function () {
        if (2 !== t.shape.length)
          throw new Error(
            'qr2d() requires a 2D Tensor, but got a ' +
              t.shape.length +
              'D Tensor.'
          )
        for (
          var n = t.shape[0],
            r = t.shape[1],
            o = Cr(n),
            a = t.clone(),
            i = Bn([[1]], [1, 1]),
            s = i.clone(),
            u = n >= r ? r : n,
            c = function (t) {
              var e,
                u = a,
                c = s,
                l = o
              ;(e = Lt.tidy(function () {
                var e = a.slice([t, t], [n - t, 1]),
                  u = e.norm(),
                  c = a.slice([t, t], [1, 1]),
                  l = Bn([[-1]]).where(c.greater(0), Bn([[1]])),
                  h = c.sub(l.mul(u)),
                  f = e.div(h)
                s =
                  1 === f.shape[0]
                    ? i.clone()
                    : i.concat(f.slice([1, 0], [f.shape[0] - 1, f.shape[1]]), 0)
                var d = l.matMul(h).div(u).neg(),
                  p = a.slice([t, 0], [n - t, r]),
                  v = d.mul(s)
                if (0 === t) a = p.sub(v.matMul(s.transpose().matMul(p)))
                else {
                  var m = p.sub(v.matMul(s.transpose().matMul(p)))
                  a = a.slice([0, 0], [t, r]).concat(m, 0)
                }
                var g = o.slice([0, t], [n, o.shape[1] - t])
                if (0 === t) o = g.sub(g.matMul(s).matMul(v.transpose()))
                else {
                  var y = g.sub(g.matMul(s).matMul(v.transpose()))
                  o = o.slice([0, 0], [n, t]).concat(y, 1)
                }
                return [s, a, o]
              })),
                (s = e[0]),
                (a = e[1]),
                (o = e[2]),
                tn([u, c, l])
            },
            l = 0;
          l < u;
          ++l
        )
          c(l)
        return (
          !e &&
            n > r &&
            ((o = o.slice([0, 0], [n, r])), (a = a.slice([0, 0], [r, r]))),
          [o, a]
        )
      })
    )
  }
  var Ch = An({
      bandPart_: function (t, e, n) {
        if (e % 1 != 0)
          throw new Error(
            'bandPart(): numLower must be an integer, got ' + e + '.'
          )
        if (n % 1 != 0)
          throw new Error(
            'bandPart(): numUpper must be an integer, got ' + n + '.'
          )
        var r = mn(t, 'a', 'bandPart')
        if (r.rank < 2)
          throw new Error(
            'bandPart(): Rank must be at least 2, got ' + r.rank + '.'
          )
        var o = r.shape,
          a = r.shape.slice(-2),
          i = a[0],
          s = a[1]
        if (!(e <= i))
          throw new Error(
            'bandPart(): numLower (' +
              e +
              ') must not be greater than the number of rows (' +
              i +
              ').'
          )
        if (!(n <= s))
          throw new Error(
            'bandPart(): numUpper (' +
              n +
              ') must not be greater than the number of columns (' +
              s +
              ').'
          )
        e < 0 && (e = i), n < 0 && (n = s)
        var u = Kn(0, i, 1, 'int32').reshape([-1, 1]),
          c = Kn(0, s, 1, 'int32'),
          l = Cc(u, c),
          h = Qu(l.lessEqual(On(+e, 'int32')), l.greaterEqual(On(-n, 'int32'))),
          f = Gn([i, s], r.dtype)
        return Pr(
          Ur(r.reshape([-1, i, s])).map(function (t) {
            return ec(h, t, f)
          })
        ).reshape(o)
      },
    }),
    Eh = An({
      gramSchmidt_: function (t) {
        var e
        if (Array.isArray(t)) {
          ;(e = !1),
            C(null != t && t.length > 0, function () {
              return 'Gram-Schmidt process: input must not be null, undefined, or empty'
            })
          for (
            var n = t[0].shape[0],
              r = function (e) {
                C(t[e].shape[0] === n, function () {
                  return (
                    'Gram-Schmidt: Non-unique lengths found in the input vectors: (' +
                    t[e].shape[0] +
                    ' vs. ' +
                    n +
                    ')'
                  )
                })
              },
              o = 1;
            o < t.length;
            ++o
          )
            r(o)
        } else
          (e = !0),
            (t = tr(t, t.shape[0], 0).map(function (t) {
              return Br(t, [0])
            }))
        C(t.length <= t[0].shape[0], function () {
          return (
            'Gram-Schmidt: Number of vectors (' +
            t.length +
            ') exceeds number of dimensions (' +
            t[0].shape[0] +
            ').'
          )
        })
        var a = [],
          i = t,
          s = function (t) {
            a.push(
              Lt.tidy(function () {
                var e = i[t]
                if (t > 0)
                  for (var n = 0; n < t; ++n) {
                    var r = Nl(a[n].mulStrict(e)).mul(a[n])
                    e = e.sub(r)
                  }
                return e.div(Vl(e, 'euclidean'))
              })
            )
          }
        for (o = 0; o < t.length; ++o) s(o)
        return e ? Pr(a, 0) : a
      },
    }),
    Rh = An({
      qr_: function (t, e) {
        if ((void 0 === e && (e = !1), t.rank < 2))
          throw new Error(
            'qr() requires input tensor to have a rank >= 2, but got rank ' +
              t.rank
          )
        if (2 === t.rank) return wh(t, e)
        var n = t.shape.slice(0, t.shape.length - 2).reduce(function (t, e) {
            return t * e
          }),
          r = Ur(
            t.reshape([
              n,
              t.shape[t.shape.length - 2],
              t.shape[t.shape.length - 1],
            ]),
            0
          ),
          o = [],
          a = []
        return (
          r.forEach(function (t) {
            var n = wh(t, e),
              r = n[0],
              i = n[1]
            o.push(r), a.push(i)
          }),
          [Pr(o, 0).reshape(t.shape), Pr(a, 0).reshape(t.shape)]
        )
      },
    }),
    Ih = Object.freeze({ bandPart: Ch, gramSchmidt: Eh, qr: Rh })
  function kh(t, e, n, r, o, a) {
    null == r && (r = 0.5),
      null == o && (o = Number.NEGATIVE_INFINITY),
      null == a && (a = 0)
    var i = t.shape[0]
    return (
      (n = Math.min(n, i)),
      C(0 <= r && r <= 1, function () {
        return "iouThreshold must be in [0, 1], but was '" + r + "'"
      }),
      C(2 === t.rank, function () {
        return "boxes must be a 2D tensor, but was of rank '" + t.rank + "'"
      }),
      C(4 === t.shape[1], function () {
        return 'boxes must have 4 columns, but 2nd dimension was ' + t.shape[1]
      }),
      C(1 === e.rank, function () {
        return 'scores must be a 1D tensor'
      }),
      C(e.shape[0] === i, function () {
        return (
          'scores has incompatible shape with boxes. Expected ' +
          i +
          ', but was ' +
          e.shape[0]
        )
      }),
      C(0 <= a && a <= 1, function () {
        return "softNmsSigma must be in [0, 1], but was '" + a + "'"
      }),
      { maxOutputSize: n, iouThreshold: r, scoreThreshold: o, softNmsSigma: a }
    )
  }
  var Sh = An({
      resizeBilinear_: function (t, e, n) {
        void 0 === n && (n = !1)
        var r = mn(t, 'images', 'resizeBilinear')
        C(3 === r.rank || 4 === r.rank, function () {
          return (
            'Error in resizeBilinear: x must be rank 3 or 4, but got rank ' +
            r.rank +
            '.'
          )
        }),
          C(2 === e.length, function () {
            return (
              'Error in resizeBilinear: new shape must 2D, but got shape ' +
              e +
              '.'
            )
          })
        var o = r,
          a = !1
        3 === r.rank &&
          ((a = !0), (o = r.as4D(1, r.shape[0], r.shape[1], r.shape[2])))
        var i = e[0],
          s = e[1],
          u = Lt.runKernelFunc(
            function (t, e) {
              return e([o]), t.resizeBilinear(o, i, s, n)
            },
            { x: o },
            function (t, e) {
              return {
                x: function () {
                  return Lt.runKernelFunc(function (r) {
                    return r.resizeBilinearBackprop(t, e[0], n)
                  }, {})
                },
              }
            },
            'ResizeBilinear',
            { alignCorners: n, newHeight: i, newWidth: s }
          )
        return a ? u.as3D(u.shape[1], u.shape[2], u.shape[3]) : u
      },
    }),
    Ah = An({
      resizeNearestNeighbor_: function (t, e, n) {
        void 0 === n && (n = !1)
        var r = mn(t, 'images', 'resizeNearestNeighbor')
        C(3 === r.rank || 4 === r.rank, function () {
          return (
            'Error in resizeNearestNeighbor: x must be rank 3 or 4, but got rank ' +
            r.rank +
            '.'
          )
        }),
          C(2 === e.length, function () {
            return (
              'Error in resizeNearestNeighbor: new shape must 2D, but got shape ' +
              e +
              '.'
            )
          }),
          C('float32' === r.dtype || 'int32' === r.dtype, function () {
            return '`images` must have `int32` or `float32` as dtype'
          })
        var o = r,
          a = !1
        3 === r.rank &&
          ((a = !0), (o = r.as4D(1, r.shape[0], r.shape[1], r.shape[2])))
        var i = e[0],
          s = e[1],
          u = Lt.runKernelFunc(
            function (t, e) {
              return e([o]), t.resizeNearestNeighbor(o, i, s, n)
            },
            { batchImages: o },
            function (t, e) {
              return {
                batchImages: function () {
                  return Lt.runKernelFunc(function (r) {
                    return r.resizeNearestNeighborBackprop(t, e[0], n)
                  }, {})
                },
              }
            }
          )
        return a ? u.as3D(u.shape[1], u.shape[2], u.shape[3]) : u
      },
    }),
    Dh = An({
      nonMaxSuppression_: function (t, e, n, r, o) {
        void 0 === r && (r = 0.5),
          void 0 === o && (o = Number.NEGATIVE_INFINITY)
        var a = mn(t, 'boxes', 'nonMaxSuppression'),
          i = mn(e, 'scores', 'nonMaxSuppression'),
          s = kh(a, i, n, r, o)
        ;(n = s.maxOutputSize), (r = s.iouThreshold), (o = s.scoreThreshold)
        var u = { maxOutputSize: n, iouThreshold: r, scoreThreshold: o }
        return Lt.runKernelFunc(
          function (t) {
            return t.nonMaxSuppression(a, i, n, r, o)
          },
          { boxes: a, scores: i },
          null,
          'NonMaxSuppressionV3',
          u
        )
      },
    }),
    Th = function (t, e, o, a, i) {
      return (
        void 0 === a && (a = 0.5),
        void 0 === i && (i = Number.NEGATIVE_INFINITY),
        n(this, void 0, void 0, function () {
          var n, s, u, c, l, h, f
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                return (
                  (n = mn(t, 'boxes', 'nonMaxSuppressionAsync')),
                  (s = mn(e, 'scores', 'nonMaxSuppressionAsync')),
                  (u = kh(n, s, o, a, i)),
                  (o = u.maxOutputSize),
                  (a = u.iouThreshold),
                  (i = u.scoreThreshold),
                  [4, Promise.all([n.data(), s.data()])]
                )
              case 1:
                return (
                  (c = r.sent()),
                  (l = c[0]),
                  (h = c[1]),
                  (f = jo(l, h, o, a, i)),
                  n !== t && n.dispose(),
                  s !== e && s.dispose(),
                  [2, f]
                )
            }
          })
        })
      )
    },
    Nh = An({
      nonMaxSuppressionWithScore_: function (t, e, n, r, o, a) {
        void 0 === r && (r = 0.5),
          void 0 === o && (o = Number.NEGATIVE_INFINITY),
          void 0 === a && (a = 0)
        var i = mn(t, 'boxes', 'nonMaxSuppression'),
          s = mn(e, 'scores', 'nonMaxSuppression'),
          u = kh(i, s, n, r, o, a),
          c = {
            maxOutputSize: (n = u.maxOutputSize),
            iouThreshold: (r = u.iouThreshold),
            scoreThreshold: (o = u.scoreThreshold),
            softNmsSigma: (a = u.softNmsSigma),
          },
          l = Lt.runKernel('NonMaxSuppressionV5', { boxes: i, scores: s }, c)
        return { selectedIndices: l[0], selectedScores: l[1] }
      },
    }),
    Fh = function (t, e, o, a, i, s) {
      return (
        void 0 === a && (a = 0.5),
        void 0 === i && (i = Number.NEGATIVE_INFINITY),
        void 0 === s && (s = 0),
        n(this, void 0, void 0, function () {
          var n, u, c, l, h, f, d
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                return (
                  (n = mn(t, 'boxes', 'nonMaxSuppressionAsync')),
                  (u = mn(e, 'scores', 'nonMaxSuppressionAsync')),
                  (c = kh(n, u, o, a, i, s)),
                  (o = c.maxOutputSize),
                  (a = c.iouThreshold),
                  (i = c.scoreThreshold),
                  (s = c.softNmsSigma),
                  [4, Promise.all([n.data(), u.data()])]
                )
              case 1:
                return (
                  (l = r.sent()),
                  (h = l[0]),
                  (f = l[1]),
                  (d = Xo(h, f, o, a, i, s)),
                  n !== t && n.dispose(),
                  u !== e && u.dispose(),
                  [2, d]
                )
            }
          })
        })
      )
    },
    _h = An({
      cropAndResize_: function (t, e, n, r, o, a) {
        var i = mn(t, 'image', 'cropAndResize'),
          s = mn(e, 'boxes', 'cropAndResize', 'float32'),
          u = mn(n, 'boxInd', 'cropAndResize', 'int32')
        ;(o = o || 'bilinear'), (a = a || 0)
        var c = s.shape[0]
        return (
          C(4 === i.rank, function () {
            return (
              'Error in cropAndResize: image must be rank 4,but got rank ' +
              i.rank +
              '.'
            )
          }),
          C(2 === s.rank && 4 === s.shape[1], function () {
            return (
              'Error in cropAndResize: boxes must be have size [' +
              c +
              ',4] but had shape ' +
              s.shape +
              '.'
            )
          }),
          C(1 === u.rank && u.shape[0] === c, function () {
            return (
              'Error in cropAndResize: boxInd must be have size [' +
              c +
              '] but had shape ' +
              s.shape +
              '.'
            )
          }),
          C(2 === r.length, function () {
            return (
              'Error in cropAndResize: cropSize must be of length 2, but got length ' +
              r.length +
              '.'
            )
          }),
          C(r[0] >= 1 && r[1] >= 1, function () {
            return 'cropSize must be atleast [1,1], but was ' + r
          }),
          C('bilinear' === o || 'nearest' === o, function () {
            return 'method must be bilinear or nearest, but was ' + o
          }),
          Lt.runKernelFunc(
            function (t, e) {
              return t.cropAndResize(i, s, u, r, o, a)
            },
            { images: i, boxes: s, boxInd: u },
            null,
            'CropAndResize',
            { method: o, extrapolationValue: a, cropSize: r }
          )
        )
      },
    }),
    Oh = Object.freeze({
      resizeBilinear: Sh,
      resizeNearestNeighbor: Ah,
      nonMaxSuppression: Dh,
      nonMaxSuppressionAsync: Th,
      nonMaxSuppressionWithScore: Nh,
      nonMaxSuppressionWithScoreAsync: Fh,
      cropAndResize: _h,
    }),
    Mh = function (t, e) {
      return !(t > 0) || 'linear' === e
    },
    Bh = function (t, e, n) {
      if (null == n || 'linear' === n) return t
      if ('relu' === n) return t.mul(e.step())
      throw new Error(
        'Gradient for activation ' + n + ' has not been implemented yet.'
      )
    },
    Ph = function (t, e) {
      var n = e,
        r = Eo(t.shape, e.shape)
      return r.length > 0 && (n = n.sum(r)), n.reshape(t.shape)
    },
    Lh = function (t, e, n) {
      if ('linear' === e) return t
      if ('relu' === e) return Bl(t)
      if ('elu' === e) return _l(t)
      if ('relu6' === e) return Pl(t)
      if ('prelu' === e) return Ml(t, n)
      throw new Error('Unknown fused activation ' + e + '.')
    }
  var Wh = An({
      fusedMatMul_: function (t) {
        var e,
          n = t.a,
          r = t.b,
          o = t.transposeA,
          a = void 0 !== o && o,
          i = t.transposeB,
          s = void 0 !== i && i,
          u = t.bias,
          c = t.activation,
          l = void 0 === c ? 'linear' : c,
          h = t.preluActivationWeights
        if (!1 === Mh(Lt.state.gradientDepth, l)) {
          var f = el(n, r, a, s)
          return null != u && (f = rc(f, u)), Lh(f, l, h)
        }
        var d = mn(n, 'a', 'fused matMul'),
          p = mn(r, 'b', 'fused matMul')
        ;(e = Nt(d, p)), (d = e[0]), (p = e[1])
        var v = a ? d.shape[d.rank - 2] : d.shape[d.rank - 1],
          m = s ? p.shape[p.rank - 1] : p.shape[p.rank - 2],
          g = a ? d.shape[d.rank - 1] : d.shape[d.rank - 2],
          y = s ? p.shape[p.rank - 2] : p.shape[p.rank - 1],
          x = d.shape.slice(0, -2),
          b = p.shape.slice(0, -2),
          w = k(x),
          E = k(b)
        C(d.rank >= 2 && p.rank >= 2 && d.rank === p.rank, function () {
          return (
            'Error in fused matMul: inputs must have the same rank of at least 2, got ranks ' +
            d.rank +
            ' and ' +
            p.rank +
            '.'
          )
        }),
          C(S(x, b), function () {
            return (
              'Error in fused matMul: outer dimensions (' +
              x +
              ') and (' +
              b +
              ') of Tensors with shapes ' +
              d.shape +
              ' and ' +
              p.shape +
              ' must match.'
            )
          }),
          C(v === m, function () {
            return (
              'Error in fused matMul: inner shapes (' +
              v +
              ') and (' +
              m +
              ') of Tensors with shapes ' +
              d.shape +
              ' and ' +
              p.shape +
              ' and transposeA=' +
              a +
              ' and transposeB=' +
              s +
              ' must match.'
            )
          })
        var R,
          I,
          A = d.shape.slice(0, -2).concat([g, y]),
          D = a ? d.as3D(w, v, g) : d.as3D(w, g, v),
          T = s ? p.as3D(E, y, m) : p.as3D(E, m, y)
        null != u &&
          Ro(A, (R = Nt((R = mn(u, 'bias', 'fused matMul')), d)[0]).shape),
          null != h && (I = mn(h, 'prelu weights', 'fused matMul'))
        var N = { a: D, b: T }
        null != u && (N.bias = R), null != h && (N.preluActivationWeights = I)
        var F = [D, T]
        return Lt.runKernelFunc(
          function (t, e) {
            var n = t.fusedBatchMatMul({
              a: D,
              b: T,
              transposeA: a,
              transposeB: s,
              bias: R,
              activation: l,
              preluActivationWeights: I,
            })
            return e([D, T, n]), n
          },
          N,
          function (t, e) {
            var n = e[0],
              r = e[1],
              o = e[2],
              i = Bh(t, o, l),
              c = {}
            return (
              null != u &&
                (c = {
                  bias: function () {
                    return Ph(R, i)
                  },
                }),
              a || s
                ? !a && s
                  ? Object.assign(
                      {
                        a: function () {
                          return i.matMul(r, !1, !1)
                        },
                        b: function () {
                          return i.matMul(n, !0, !1)
                        },
                      },
                      c
                    )
                  : a && !s
                  ? Object.assign(
                      {
                        a: function () {
                          return r.matMul(i, !1, !0)
                        },
                        b: function () {
                          return n.matMul(i, !1, !1)
                        },
                      },
                      c
                    )
                  : Object.assign(
                      {
                        a: function () {
                          return r.matMul(i, !0, !0)
                        },
                        b: function () {
                          return i.matMul(n, !0, !0)
                        },
                      },
                      c
                    )
                : Object.assign(
                    {
                      a: function () {
                        return i.matMul(r, !1, !0)
                      },
                      b: function () {
                        return n.matMul(i, !0, !1)
                      },
                    },
                    c
                  )
            )
          },
          '_FusedMatMul',
          { transposeA: a, transposeB: s, activation: l },
          F,
          [!0]
        ).reshape(A)
      },
    }),
    Uh = An({
      fusedConv2d_: function (t) {
        var e = t.x,
          n = t.filter,
          r = t.strides,
          o = t.pad,
          a = t.dataFormat,
          i = void 0 === a ? 'NHWC' : a,
          s = t.dilations,
          u = void 0 === s ? [1, 1] : s,
          c = t.dimRoundingMode,
          l = t.bias,
          h = t.activation,
          f = void 0 === h ? 'linear' : h,
          d = t.preluActivationWeights
        if (((f = f || 'linear'), !1 === Mh(Lt.state.gradientDepth, f))) {
          var p = qc(e, n, r, o, i, u, c)
          return null != l && (p = rc(p, l)), Lh(p, f, d)
        }
        var v = mn(e, 'x', 'conv2d'),
          m = mn(n, 'filter', 'conv2d'),
          g = v,
          y = !1
        3 === v.rank &&
          ((y = !0), (g = v.as4D(1, v.shape[0], v.shape[1], v.shape[2]))),
          C(4 === g.rank, function () {
            return (
              'Error in fused conv2d: input must be rank 4, but got rank ' +
              g.rank +
              '.'
            )
          }),
          C(4 === m.rank, function () {
            return (
              'Error in fused conv2d: filter must be rank 4, but got rank ' +
              m.rank +
              '.'
            )
          }),
          null != c &&
            C(A(o), function () {
              return (
                'Error in fused conv2d: pad must be an integer when using, dimRoundingMode ' +
                c +
                ' but got pad ' +
                o +
                '.'
              )
            }),
          C(g.shape[3] === m.shape[2], function () {
            return (
              'Error in conv2d: depth of input (' +
              g.shape[3] +
              ') must match input depth for filter ' +
              m.shape[2] +
              '.'
            )
          }),
          C(Mo(r, u), function () {
            return (
              'Error in conv2D: Either strides or dilations must be 1. Got strides ' +
              r +
              " and dilations '" +
              u +
              "'"
            )
          }),
          C('NHWC' === i, function () {
            return (
              'Error in conv2d: got dataFormat of ' +
              i +
              ' but only NHWC is currently supported.'
            )
          })
        var x,
          b,
          w = So(g.shape, m.shape, r, u, o, c)
        null != l &&
          ((x = Nt((x = mn(l, 'bias', 'fused conv2d')), v)[0]),
          Ro(w.outShape, x.shape)),
          null != d && (b = mn(d, 'prelu weights', 'fused conv2d'))
        var E = { x: g, filter: m }
        null != l && (E.bias = x), null != d && (E.preluActivationWeights = b)
        var R = [m, g],
          I = Lt.runKernelFunc(
            function (t, e) {
              var n = t.fusedConv2d({
                input: g,
                filter: m,
                convInfo: w,
                bias: x,
                activation: f,
                preluActivationWeights: b,
              })
              return e([m, g, n]), n
            },
            E,
            function (t, e) {
              var n = e,
                a = n[0],
                i = n[1],
                s = n[2],
                c = Bh(t, s, f)
              C(Oo(u), function () {
                return (
                  "Error in gradient of fused conv2D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '" +
                  u +
                  "'"
                )
              })
              var h = {}
              return (
                null != l &&
                  (h = {
                    bias: function () {
                      return Ph(x, c)
                    },
                  }),
                Object.assign(
                  {
                    x: function () {
                      return Xc(i.shape, c, a, r, o)
                    },
                    filter: function () {
                      return jc(i, c, a.shape, r, o)
                    },
                  },
                  h
                )
              )
            },
            'FusedConv2D',
            { convInfo: w, activation: f },
            R,
            [!0]
          )
        return y ? I.as3D(I.shape[1], I.shape[2], I.shape[3]) : I
      },
    }),
    Vh = An({
      fusedDepthwiseConv2d_: function (t) {
        var e = t.x,
          n = t.filter,
          r = t.strides,
          o = t.pad,
          a = t.dataFormat,
          i = void 0 === a ? 'NHWC' : a,
          s = t.dilations,
          u = void 0 === s ? [1, 1] : s,
          c = t.dimRoundingMode,
          l = t.bias,
          h = t.activation,
          f = void 0 === h ? 'linear' : h,
          d = t.preluActivationWeights
        if (!1 === Mh(Lt.state.gradientDepth, f)) {
          var p = Yc(e, n, r, o, i, u, c)
          return null != l && (p = rc(p, l)), Lh(p, f, d)
        }
        var v = mn(e, 'x', 'depthwiseConv2d'),
          m = mn(n, 'filter', 'depthwiseConv2d'),
          g = v,
          y = !1
        3 === v.rank &&
          ((y = !0), (g = v.as4D(1, v.shape[0], v.shape[1], v.shape[2]))),
          C(4 === g.rank, function () {
            return (
              'Error in fused depthwiseConv2d: input must be rank 4, but got rank ' +
              g.rank +
              '.'
            )
          }),
          C(4 === m.rank, function () {
            return (
              'Error in fused depthwiseConv2d: filter must be rank 4, but got rank ' +
              m.rank +
              '.'
            )
          }),
          C(g.shape[3] === m.shape[2], function () {
            return (
              'Error in fused depthwiseConv2d: number of input channels (' +
              g.shape[3] +
              ') must match the inChannels dimension in filter ' +
              m.shape[2] +
              '.'
            )
          }),
          null == u && (u = [1, 1]),
          C(Mo(r, u), function () {
            return (
              'Error in fused depthwiseConv2d: Either strides or dilations must be 1. Got strides ' +
              r +
              " and dilations '" +
              u +
              "'"
            )
          }),
          null != c &&
            C(A(o), function () {
              return (
                'Error in fused depthwiseConv2d: pad must be an integer when using dimRoundingMode ' +
                c +
                ' but got pad ' +
                o +
                '.'
              )
            })
        var x,
          b,
          w = So(g.shape, m.shape, r, u, o, c, !0)
        null != l &&
          ((x = Nt((x = mn(l, 'bias', 'fused conv2d')), v)[0]),
          Ro(w.outShape, x.shape)),
          null != d && (b = mn(d, 'prelu weights', 'fused depthwiseConv2d'))
        var E = { x: g, filter: m }
        null != l && (E.bias = x), null != d && (E.preluActivationWeights = b)
        var R = [m, g],
          I = Lt.runKernelFunc(
            function (t, e) {
              var n = t.fusedDepthwiseConv2D({
                input: g,
                filter: m,
                convInfo: w,
                bias: x,
                activation: f,
                preluActivationWeights: b,
              })
              return e([m, g, n]), n
            },
            E,
            function (t, e) {
              C(Oo(u), function () {
                return (
                  "Error in gradient of fused depthwiseConv2d: dilation rates greater than 1 are not yet supported. Got dilations '" +
                  u +
                  "'"
                )
              })
              var n = e[0],
                r = e[1],
                o = e[2],
                a = Bh(t, o, f),
                i = {}
              return (
                null != l &&
                  (i = {
                    bias: function () {
                      return Ph(x, a)
                    },
                  }),
                Object.assign(
                  {
                    x: function () {
                      return $c(r.shape, a, n, w)
                    },
                    filter: function () {
                      return Qc(r, a, n.shape, w)
                    },
                  },
                  i
                )
              )
            },
            'FusedDepthwiseConv2D',
            { convInfo: w, activation: f },
            R,
            [!0]
          )
        return y ? I.as3D(I.shape[1], I.shape[2], I.shape[3]) : I
      },
    }),
    zh = Object.freeze({ matMul: Wh, conv2d: Uh, depthwiseConv2d: Vh }),
    Gh = Object.freeze({
      image: Oh,
      linalg: Ih,
      losses: bh,
      spectral: Jl,
      fused: zh,
      signal: uh,
      square: tu,
      squaredDifference: nu,
      conv1d: Hc,
      conv2d: qc,
      conv3d: Kc,
      depthwiseConv2d: Yc,
      separableConv2d: Jc,
      conv2dTranspose: Zc,
      conv3dTranspose: tl,
      op: An,
      batchNormalization2d: Gu,
      batchNormalization3d: Hu,
      batchNormalization4d: qu,
      batchNormalization: Ku,
      batchNorm: ju,
      batchNorm2d: Xu,
      batchNorm3d: Yu,
      batchNorm4d: $u,
      booleanMaskAsync: Uc,
      complex: Dn,
      real: Tn,
      imag: Nn,
      concat: Yn,
      concat1d: $n,
      concat2d: Qn,
      concat3d: Jn,
      concat4d: Zn,
      split: tr,
      matMul: el,
      dot: nl,
      outerProduct: rl,
      reverse: ol,
      reverse1d: al,
      reverse2d: il,
      reverse3d: sl,
      reverse4d: ul,
      maxPool: hl,
      avgPool: fl,
      pool: dl,
      maxPool3d: pl,
      avgPool3d: vl,
      slice: ml,
      slice1d: gl,
      slice2d: yl,
      slice3d: xl,
      slice4d: bl,
      abs: ru,
      acos: ou,
      acosh: au,
      asin: iu,
      asinh: su,
      atan: uu,
      atanh: cu,
      ceil: lu,
      clipByValue: hu,
      cos: fu,
      cosh: du,
      erf: pu,
      exp: vu,
      expm1: mu,
      floor: gu,
      log: yu,
      log1p: xu,
      logSigmoid: bu,
      neg: wu,
      reciprocal: Cu,
      round: Eu,
      rsqrt: Ru,
      sigmoid: Iu,
      sign: ku,
      isNaN: Su,
      isInf: Au,
      isFinite: Du,
      sin: Tu,
      sinh: Nu,
      softplus: Fu,
      sqrt: _u,
      step: Ou,
      tan: Mu,
      tanh: Bu,
      all: Cl,
      any: El,
      argMax: Rl,
      argMin: Il,
      logSumExp: kl,
      max: Sl,
      mean: Al,
      min: Dl,
      moments: Tl,
      sum: Nl,
      prod: Fl,
      equal: Rc,
      equalStrict: Ic,
      greater: kc,
      greaterEqual: Sc,
      greaterEqualStrict: Ac,
      greaterStrict: Dc,
      less: Tc,
      lessEqual: Nc,
      lessEqualStrict: Fc,
      lessStrict: _c,
      notEqual: Oc,
      notEqualStrict: Mc,
      add: rc,
      addN: oc,
      addStrict: ac,
      atan2: ic,
      div: sc,
      divNoNan: uc,
      divStrict: cc,
      floorDiv: lc,
      maximum: hc,
      maximumStrict: fc,
      minimum: dc,
      minimumStrict: pc,
      mod: vc,
      modStrict: mc,
      mul: gc,
      mulStrict: yc,
      pow: xc,
      powStrict: bc,
      squaredDifferenceStrict: wc,
      sub: Cc,
      subStrict: Ec,
      elu: _l,
      leakyRelu: Ol,
      prelu: Ml,
      relu: Bl,
      relu6: Pl,
      selu: Ll,
      logicalAnd: Qu,
      logicalNot: Ju,
      logicalOr: Zu,
      logicalXor: tc,
      where: ec,
      whereAsync: nc,
      buffer: dr,
      print: pr,
      batchToSpaceND: vr,
      broadcastTo: mr,
      cast: gr,
      clone: yr,
      cumsum: xr,
      depthToSpace: br,
      expandDims: wr,
      eye: Cr,
      multinomial: Er,
      oneHot: Rr,
      pad: Ir,
      pad1d: kr,
      pad2d: Sr,
      pad3d: Ar,
      pad4d: Dr,
      rand: Tr,
      randomNormal: Nr,
      randomGamma: Fr,
      randomUniform: _r,
      reshape: Or,
      spaceToBatchND: Mr,
      squeeze: Br,
      stack: Pr,
      tile: Lr,
      truncatedNormal: Wr,
      unstack: Ur,
      setdiff1dAsync: Vr,
      fill: Hn,
      linspace: qn,
      ones: zn,
      range: Kn,
      scalar: On,
      tensor: Fn,
      tensor1d: Mn,
      tensor2d: Bn,
      tensor3d: Pn,
      tensor4d: Ln,
      tensor5d: Wn,
      tensor6d: Un,
      variable: Vn,
      zeros: Gn,
      onesLike: jn,
      zerosLike: Xn,
      transpose: Wl,
      softmax: go,
      logSoftmax: yo,
      localResponseNormalization: Ul,
      norm: Vl,
      gather: Lc,
      unsortedSegmentSum: Wc,
      basicLSTMCell: zl,
      multiRNNCell: Gl,
      movingAverage: Hl,
      stridedSlice: ql,
      topk: Kl,
      scatterND: jl,
      fft: Xl,
      ifft: Yl,
      rfft: $l,
      irfft: Ql,
      sparseToDense: Zl,
      gatherND: th,
      diag: eh,
      dropout: nh,
      hannWindow: oh,
      hammingWindow: ah,
      frame: ih,
      stft: sh,
      inTopKAsync: lh,
    })
  function Hh(t, e) {
    Array.isArray(t) || (t = [t]),
      t.forEach(function (t) {
        null != t &&
          C('complex64' !== t.dtype, function () {
            return e + ' does not support complex64 tensors.'
          })
      })
  }
  function qh(t, e, n, r) {
    if ('linear' === n) return t.linear(e)
    if ('relu' === n) return t.relu(e)
    if ('elu' === n) return t.elu(e)
    if ('relu6' === n) return t.relu6(e)
    if ('prelu' === n) return t.prelu(e, r)
    throw new Error(
      'Activation ' + n + ' has not been implemented for the CPU backend.'
    )
  }
  var Kh = (function (t) {
    function o() {
      var e = t.call(this) || this
      return (e.blockSize = 48), (e.firstUse = !0), (e.data = new xo(e, Lt)), e
    }
    return (
      e(o, t),
      (o.prototype.write = function (t, e, n) {
        this.firstUse &&
          ((this.firstUse = !1),
          i().get('IS_NODE') &&
            dn(
              "\n============================\nHi there 👋. Looks like you are running TensorFlow.js in Node.js. To speed things up dramatically, install our node backend, which binds to TensorFlow C++, by running npm i @tensorflow/tfjs-node, or npm i @tensorflow/tfjs-node-gpu if you have CUDA. Then call require('@tensorflow/tfjs-node'); (-gpu suffix for CUDA) at the start of your program. Visit https://github.com/tensorflow/tfjs-node for more details.\n============================"
            ))
        var r = {}
        return this.data.set(r, { values: t, dtype: n }), r
      }),
      (o.prototype.move = function (t, e, n, r) {
        this.data.set(t, { values: e, dtype: r })
      }),
      (o.prototype.numDataIds = function () {
        return this.data.numDataIds()
      }),
      (o.prototype.read = function (t) {
        return n(this, void 0, void 0, function () {
          return r(this, function (e) {
            return [2, this.readSync(t)]
          })
        })
      }),
      (o.prototype.readSync = function (t) {
        var e = this.data.get(t),
          n = e.dtype,
          r = e.complexTensors
        return 'complex64' === n
          ? Vo(this.readSync(r.real.dataId), this.readSync(r.imag.dataId))
          : this.data.get(t).values
      }),
      (o.prototype.bufferSync = function (t) {
        var e = this.readSync(t.dataId),
          n = e
        if ('string' === t.dtype)
          try {
            n = e.map(function (t) {
              return ot(t)
            })
          } catch (t) {
            throw new Error('Failed to decode encoded string bytes into utf-8')
          }
        return dr(t.shape, t.dtype, n)
      }),
      (o.prototype.makeOutput = function (t, e, n) {
        var r = this.write(t, e, n)
        return Lt.makeTensorFromDataId(r, e, n, this)
      }),
      (o.prototype.disposeData = function (t) {
        if (this.data.has(t)) {
          var e = this.data.get(t).complexTensors
          null != e && (e.real.dispose(), e.imag.dispose()), this.data.delete(t)
        }
      }),
      (o.prototype.time = function (t) {
        return n(this, void 0, void 0, function () {
          var e
          return r(this, function (n) {
            return (e = et()), t(), [2, { kernelMs: et() - e }]
          })
        })
      }),
      (o.prototype.memory = function () {
        return {
          unreliable: !0,
          reasons: [
            'The reported memory is an upper bound. Due to automatic garbage collection, the true allocated memory may be less.',
          ],
        }
      }),
      (o.prototype.complex = function (t, e) {
        var n = this.makeOutput(null, t.shape, 'complex64')
        return (
          (this.data.get(n.dataId).complexTensors = {
            real: Lt.keep(t.clone()),
            imag: Lt.keep(e.clone()),
          }),
          n
        )
      }),
      (o.prototype.real = function (t) {
        return this.data.get(t.dataId).complexTensors.real.clone()
      }),
      (o.prototype.imag = function (t) {
        return this.data.get(t.dataId).complexTensors.imag.clone()
      }),
      (o.prototype.slice = function (t, e, n) {
        if ((Hh(t, 'slice'), io(t.shape, e, n))) {
          var r = so(e, t.strides),
            o = k(n)
          return Fn(this.readSync(t.dataId).subarray(r, r + o), n, t.dtype)
        }
        for (
          var a = dr(n, t.dtype), i = this.bufferSync(t), s = 0;
          s < a.size;
          ++s
        ) {
          var u = a.indexToLoc(s).map(function (t, n) {
            return t + e[n]
          })
          a.values[s] = i.get.apply(i, u)
        }
        return a.toTensor()
      }),
      (o.prototype.stridedSlice = function (t, e, n, r) {
        Hh(t, 'stridedSlice')
        var o = ro(e, n, r)
        if (
          o.some(function (t) {
            return 0 === t
          })
        )
          return Fn([], o)
        for (
          var a = dr(o, t.dtype), i = this.bufferSync(t), s = 0;
          s < a.size;
          s++
        ) {
          for (
            var u = a.indexToLoc(s), c = new Array(u.length), l = 0;
            l < c.length;
            l++
          )
            c[l] = u[l] * r[l] + e[l]
          a.set.apply(a, [i.get.apply(i, c)].concat(u))
        }
        return a.toTensor()
      }),
      (o.prototype.diag = function (t) {
        for (
          var e = this.readSync(t.dataId),
            n = dr([t.size, t.size], t.dtype),
            r = n.values,
            o = 0;
          o < e.length;
          o++
        )
          r[o * t.size + o] = e[o]
        return n.toTensor()
      }),
      (o.prototype.unstack = function (t, e) {
        for (
          var n = t.shape[e], r = new Array(t.rank - 1), o = 0, a = 0;
          a < t.rank;
          a++
        )
          a !== e && (r[o++] = t.shape[a])
        var i = new Array(t.rank).fill(0),
          s = t.shape.slice()
        s[e] = 1
        var u = new Array(n)
        for (a = 0; a < u.length; a++)
          (i[e] = a), (u[a] = this.slice(t, i, s).reshape(r))
        return u
      }),
      (o.prototype.reverse = function (t, e) {
        Hh(t, 'reverse')
        for (
          var n = dr(t.shape, t.dtype),
            r = this.bufferSync(t),
            o = function (o) {
              var a = n.indexToLoc(o),
                i = a.slice()
              e.forEach(function (e) {
                return (i[e] = t.shape[e] - 1 - i[e])
              }),
                n.set.apply(n, [r.get.apply(r, i)].concat(a))
            },
            a = 0;
          a < n.size;
          a++
        )
          o(a)
        return n.toTensor()
      }),
      (o.prototype.concat = function (t, e) {
        var n = this
        if ('complex64' === t[0].dtype) {
          var r = t.map(function (t) {
              return Tn(t)
            }),
            o = t.map(function (t) {
              return Nn(t)
            })
          return Dn(this.concat(r, e), this.concat(o, e))
        }
        var a = t.map(function (t) {
            var n = k(t.shape.slice(e))
            return t.as2D(-1, n)
          }),
          i = Sn(
            a.map(function (t) {
              return t.shape
            }),
            1
          ),
          s = dr(i, t[0].dtype).values
        if (1 === a[0].shape[0]) {
          var u = 0
          a.forEach(function (t) {
            s.set(n.readSync(t.dataId), u), (u += t.size)
          })
        } else {
          var c = 0
          a.forEach(function (t) {
            for (
              var e = n.readSync(t.dataId), r = 0, o = 0;
              o < t.shape[0];
              ++o
            )
              for (var a = o * i[1] + c, u = 0; u < t.shape[1]; ++u)
                s[a + u] = e[r++]
            c += t.shape[1]
          })
        }
        var l = Sn(
          t.map(function (t) {
            return t.shape
          }),
          e
        )
        return Fn(s, l, t[0].dtype)
      }),
      (o.prototype.neg = function (t) {
        return Hh(t, 'neg'), this.multiply(On(-1), t)
      }),
      (o.prototype.add = function (t, e) {
        return 'complex64' === t.dtype || 'complex64' === e.dtype
          ? this.broadcastedBinaryComplexOp(
              t.cast('complex64'),
              e.cast('complex64'),
              function (t, e, n, r) {
                return { real: t + n, imag: e + r }
              }
            )
          : this.broadcastedBinaryOp(
              t,
              e,
              Dt(t.dtype, e.dtype),
              function (t, e) {
                return t + e
              }
            )
      }),
      (o.prototype.addN = function (t) {
        var e = this
        Hh(t, 'addN')
        for (
          var n = t.map(function (t) {
              return e.readSync(t.dataId)
            }),
            r = dr(t[0].shape, t[0].dtype),
            o = r.values,
            a = 0;
          a < t.length;
          a++
        )
          for (var i = n[a], s = 0; s < o.length; s++) o[s] += i[s]
        return r.toTensor()
      }),
      (o.prototype.softmax = function (t, e) {
        var n = O([e], t.shape),
          r = this.max(t, n),
          o = wn(r.shape, n),
          a = this.subtract(t, r.reshape(o)),
          i = this.exp(a),
          s = this.sum(i, n).reshape(o)
        return this.realDivide(i, s)
      }),
      (o.prototype.subtract = function (t, e) {
        return 'complex64' === t.dtype || 'complex64' === e.dtype
          ? this.broadcastedBinaryComplexOp(
              t.cast('complex64'),
              e.cast('complex64'),
              function (t, e, n, r) {
                return { real: t - n, imag: e - r }
              }
            )
          : this.broadcastedBinaryOp(
              t,
              e,
              Dt(t.dtype, e.dtype),
              function (t, e) {
                return t - e
              }
            )
      }),
      (o.prototype.pow = function (t, e) {
        return (
          Hh([t, e], 'pow'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            return Math.pow(t, e)
          })
        )
      }),
      (o.prototype.batchMatMul = function (t, e, n, r) {
        Hh([t, e], 'matMul')
        for (
          var o = n ? t.shape[1] : t.shape[2],
            a = n ? t.shape[2] : t.shape[1],
            i = r ? e.shape[1] : e.shape[2],
            s = t.shape[0],
            u = this.readSync(t.dataId),
            c = this.readSync(e.dataId),
            l = n
              ? [t.strides[0], 1, t.strides[1]]
              : [t.strides[0], t.strides[1], 1],
            h = l[0],
            f = l[1],
            d = l[2],
            p = r
              ? [1, e.strides[1], e.strides[0]]
              : [e.strides[1], 1, e.strides[0]],
            v = p[0],
            m = p[1],
            g = p[2],
            y = a * i,
            x = dr([s, a, i], t.dtype),
            b = x.values,
            w = this.blockSize,
            C = 0;
          C < s;
          C++
        )
          for (var E = 0; E < a; E += w)
            for (var R = 0; R < i; R += w)
              for (var I = 0; I < o; I += w)
                for (
                  var k = Math.min(E + w, a),
                    S = Math.min(R + w, i),
                    A = Math.min(I + w, o),
                    D = E;
                  D < k;
                  D++
                )
                  for (var T = R; T < S; T++) {
                    for (var N = 0, F = I; F < A; F++)
                      N += u[C * h + D * f + F * d] * c[F * v + T * m + C * g]
                    b[C * y + (D * i + T)] += N
                  }
        return x.toTensor()
      }),
      (o.prototype.fusedBatchMatMul = function (t) {
        var e = t.a,
          n = t.b,
          r = t.transposeA,
          o = t.transposeB,
          a = t.bias,
          i = t.activation,
          s = t.preluActivationWeights,
          u = this.batchMatMul(e, n, r, o)
        return a && (u = this.add(u, a)), i && (u = qh(this, u, i, s)), u
      }),
      (o.prototype.multiply = function (t, e) {
        return 'complex64' === t.dtype || 'complex64' === e.dtype
          ? this.broadcastedBinaryComplexOp(
              t.cast('complex64'),
              e.cast('complex64'),
              function (t, e, n, r) {
                return { real: t * n - e * r, imag: t * r + e * n }
              }
            )
          : this.broadcastedBinaryOp(
              t,
              e,
              Dt(t.dtype, e.dtype),
              function (t, e) {
                return t * e
              }
            )
      }),
      (o.prototype.realDivide = function (t, e) {
        Hh([t, e], 'realDivide')
        return this.broadcastedBinaryOp(t, e, 'float32', function (t, e) {
          return t / e
        })
      }),
      (o.prototype.floorDiv = function (t, e) {
        Hh([t, e], 'floorDiv')
        return this.broadcastedBinaryOp(t, e, 'int32', function (t, e) {
          return Math.floor(t / e)
        })
      }),
      (o.prototype.sum = function (t, e) {
        Hh(t, 'sum'), Cn('sum', e, t.rank)
        for (
          var n = bn(t.shape, e),
            r = n[0],
            o = n[1],
            a = Gn(r, Dt(t.dtype, 'int32')),
            i = k(o),
            s = this.readSync(a.dataId),
            u = this.readSync(t.dataId),
            c = 0;
          c < s.length;
          ++c
        ) {
          for (var l = c * i, h = 0, f = 0; f < i; ++f) h += u[l + f]
          s[c] = h
        }
        return a
      }),
      (o.prototype.prod = function (t, e) {
        Hh(t, 'sum')
        for (
          var n = bn(t.shape, e),
            r = n[0],
            o = n[1],
            a = Gn(r, Dt(t.dtype, 'int32')),
            i = k(o),
            s = this.readSync(a.dataId),
            u = this.readSync(t.dataId),
            c = 0;
          c < s.length;
          ++c
        ) {
          for (var l = c * i, h = 1, f = 0; f < i; ++f) h *= u[l + f]
          s[c] = h
        }
        return a
      }),
      (o.prototype.unsortedSegmentSum = function (t, e, n) {
        Hh(t, 'unsortedSegmentSum')
        for (var r = [], o = t.rank - e.rank, a = 0; a < o; ++a)
          e = e.expandDims(a + 1)
        for (a = 0; a < n; ++a) {
          var i = On(a, 'int32'),
            s = Rc(i, e).asType('float32').mul(t).sum(0)
          r.push(s)
        }
        return Pr(r)
      }),
      (o.prototype.argMin = function (t, e) {
        Hh(t, 'argMin')
        var n = [e]
        Cn('argMin', n, t.rank)
        for (
          var r = bn(t.shape, n),
            o = r[0],
            a = r[1],
            i = Gn(o, 'int32'),
            s = k(a),
            u = this.readSync(i.dataId),
            c = this.readSync(t.dataId),
            l = 0;
          l < u.length;
          ++l
        ) {
          for (var h = l * s, f = c[h], d = 0, p = 0; p < s; ++p) {
            var v = c[h + p]
            v < f && ((f = v), (d = p))
          }
          u[l] = d
        }
        return i
      }),
      (o.prototype.argMax = function (t, e) {
        Hh(t, 'argMax')
        var n = [e]
        Cn('argMax', n, t.rank)
        for (
          var r = bn(t.shape, n),
            o = r[0],
            a = r[1],
            i = Gn(o, 'int32'),
            s = k(a),
            u = this.readSync(i.dataId),
            c = this.readSync(t.dataId),
            l = 0;
          l < u.length;
          ++l
        ) {
          for (var h = l * s, f = c[h], d = 0, p = 0; p < s; ++p) {
            var v = c[h + p]
            v > f && ((f = v), (d = p))
          }
          u[l] = d
        }
        return i
      }),
      (o.prototype.cumsum = function (t, e, n, r) {
        if ((Hh(t, 'cumsum'), e !== t.rank - 1))
          throw new Error(
            'backend.cumsum in CPU expects an inner-most axis=' +
              (t.rank - 1) +
              ' but got axis=' +
              e
          )
        for (
          var o = Dt(t.dtype, 'int32'),
            a = Gn(t.shape, o),
            i = this.readSync(a.dataId),
            s = this.readSync(t.dataId),
            u = t.shape[t.rank - 1],
            c = r
              ? function (t, e) {
                  return t + u - e - 1
                }
              : function (t, e) {
                  return t + e
                },
            l = 0;
          l < s.length;
          l += u
        )
          for (var h = 0; h < u; h++) {
            var f = c(l, h)
            if (0 === h) i[f] = n ? 0 : s[f]
            else {
              var d = c(l, h - 1)
              i[f] = n ? s[d] + i[d] : s[f] + i[d]
            }
          }
        return a
      }),
      (o.prototype.equal = function (t, e) {
        return (
          Hh([t, e], 'equal'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t === e ? 1 : 0
          })
        )
      }),
      (o.prototype.notEqual = function (t, e) {
        return (
          Hh([t, e], 'notEqual'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t !== e ? 1 : 0
          })
        )
      }),
      (o.prototype.less = function (t, e) {
        return (
          Hh([t, e], 'less'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t < e ? 1 : 0
          })
        )
      }),
      (o.prototype.lessEqual = function (t, e) {
        return (
          Hh([t, e], 'lessEqual'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t <= e ? 1 : 0
          })
        )
      }),
      (o.prototype.greater = function (t, e) {
        return (
          Hh([t, e], 'greater'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t > e ? 1 : 0
          })
        )
      }),
      (o.prototype.greaterEqual = function (t, e) {
        return (
          Hh([t, e], 'greaterEqual'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t >= e ? 1 : 0
          })
        )
      }),
      (o.prototype.logicalNot = function (t) {
        Hh(t, 'logicalNot')
        for (
          var e = this.readSync(t.dataId), n = new Uint8Array(e.length), r = 0;
          r < e.length;
          ++r
        )
          n[r] = e[r] ? 0 : 1
        return this.makeOutput(n, t.shape, 'bool')
      }),
      (o.prototype.logicalAnd = function (t, e) {
        return (
          Hh([t, e], 'logicalAnd'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t && e
          })
        )
      }),
      (o.prototype.logicalOr = function (t, e) {
        return (
          Hh([t, e], 'logicalOr'),
          this.broadcastedBinaryOp(t, e, 'bool', function (t, e) {
            return t || e
          })
        )
      }),
      (o.prototype.select = function (t, e, n) {
        Hh([t, e, n], 'select')
        for (
          var r = this.readSync(t.dataId),
            o = this.readSync(e.dataId),
            a = this.readSync(n.dataId),
            i = Gn(e.shape, Dt(e.dtype, n.dtype)),
            s = this.readSync(i.dataId),
            u = 0,
            c =
              0 === t.rank || t.rank > 1 || 1 === e.rank
                ? 1
                : k(e.shape.slice(1)),
            l = 0;
          l < r.length;
          l++
        )
          for (var h = 0; h < c; h++)
            1 === r[l] ? (s[u++] = o[l]) : (s[u++] = a[l])
        return i
      }),
      (o.prototype.where = function (t) {
        Hh([t], 'where')
        var e = this.readSync(t.dataId)
        return na(t.shape, e)
      }),
      (o.prototype.topk = function (t, e, n) {
        return Hh(t, 'topk'), ea(this.readSync(t.dataId), t.shape, t.dtype, e)
      }),
      (o.prototype.min = function (t, e) {
        Hh(t, 'min'), Cn('min', e, t.rank)
        for (
          var n = bn(t.shape, e),
            r = n[0],
            o = n[1],
            a = Gn(r, t.dtype),
            i = k(o),
            s = this.readSync(a.dataId),
            u = this.readSync(t.dataId),
            c = 0;
          c < s.length;
          ++c
        ) {
          for (var l = c * i, h = u[l], f = 0; f < i; ++f) {
            var d = u[l + f]
            d < h && (h = d)
          }
          s[c] = h
        }
        return a
      }),
      (o.prototype.minimum = function (t, e) {
        return (
          Hh([t, e], 'minimum'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            return Math.min(t, e)
          })
        )
      }),
      (o.prototype.mod = function (t, e) {
        return (
          Hh([t, e], 'mod'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            var n = t % e
            return (t < 0 && e < 0) || (t >= 0 && e >= 0) ? n : (n + e) % e
          })
        )
      }),
      (o.prototype.max = function (t, e) {
        Hh(t, 'max'), Cn('max', e, t.rank)
        for (
          var n = bn(t.shape, e),
            r = n[0],
            o = n[1],
            a = Gn(r, t.dtype),
            i = k(o),
            s = this.readSync(a.dataId),
            u = this.readSync(t.dataId),
            c = 0;
          c < s.length;
          ++c
        ) {
          for (var l = c * i, h = u[l], f = 0; f < i; ++f) {
            var d = u[l + f]
            d > h && (h = d)
          }
          s[c] = h
        }
        return a
      }),
      (o.prototype.maximum = function (t, e) {
        return (
          Hh([t, e], 'maximum'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            return Math.max(t, e)
          })
        )
      }),
      (o.prototype.all = function (t, e) {
        Hh(t, 'all'), Cn('all', e, t.rank)
        for (
          var n = bn(t.shape, e),
            r = n[0],
            o = n[1],
            a = Gn(r, t.dtype),
            i = k(o),
            s = this.readSync(a.dataId),
            u = this.readSync(t.dataId),
            c = 0;
          c < s.length;
          ++c
        ) {
          for (var l = c * i, h = u[l], f = 0; f < i; ++f) {
            var d = u[l + f]
            h = h && d
          }
          s[c] = h
        }
        return a
      }),
      (o.prototype.any = function (t, e) {
        Hh(t, 'any'), Cn('any', e, t.rank)
        for (
          var n = bn(t.shape, e),
            r = n[0],
            o = n[1],
            a = Gn(r, t.dtype),
            i = k(o),
            s = this.readSync(a.dataId),
            u = this.readSync(t.dataId),
            c = 0;
          c < s.length;
          ++c
        ) {
          for (var l = c * i, h = u[l], f = 0; f < i; ++f) {
            var d = u[l + f]
            h = h || d
          }
          s[c] = h
        }
        return a
      }),
      (o.prototype.squaredDifference = function (t, e) {
        return (
          Hh([t, e], 'squaredDifference'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            var n = t - e
            return n * n
          })
        )
      }),
      (o.prototype.ceil = function (t) {
        Hh(t, 'ceil')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        )
          n[r] = Math.ceil(e[r])
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.floor = function (t) {
        Hh(t, 'floor')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        )
          n[r] = Math.floor(e[r])
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.sign = function (t) {
        Hh(t, 'x')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        )
          e[r] < 0 ? (n[r] = -1) : e[r] > 0 ? (n[r] = 1) : (n[r] = 0)
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.isNaN = function (t) {
        Hh(t, 'x')
        for (
          var e = this.readSync(t.dataId), n = new Uint8Array(e.length), r = 0;
          r < e.length;
          ++r
        )
          Number.isNaN(e[r]) && (n[r] = 1)
        return this.makeOutput(n, t.shape, 'bool')
      }),
      (o.prototype.isInf = function (t) {
        Hh(t, 'x')
        for (
          var e = this.readSync(t.dataId), n = new Uint8Array(e.length), r = 0;
          r < e.length;
          ++r
        )
          Math.abs(e[r]) === 1 / 0 && (n[r] = 1)
        return this.makeOutput(n, t.shape, 'bool')
      }),
      (o.prototype.isFinite = function (t) {
        Hh(t, 'x')
        for (
          var e = this.readSync(t.dataId), n = new Uint8Array(e.length), r = 0;
          r < e.length;
          ++r
        )
          Number.isFinite(e[r]) && (n[r] = 1)
        return this.makeOutput(n, t.shape, 'bool')
      }),
      (o.prototype.round = function (t) {
        Hh(t, 'round')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        ) {
          var o = Math.floor(e[r])
          e[r] - o < 0.5
            ? (n[r] = Math.floor(e[r]))
            : e[r] - o > 0.5
            ? (n[r] = Math.ceil(e[r]))
            : (n[r] = o % 2 == 0 ? o : o + 1)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.exp = function (t) {
        Hh(t, 'exp')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        )
          n[r] = Math.exp(e[r])
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.expm1 = function (t) {
        Hh(t, 'expm1')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        )
          n[r] = Math.expm1(e[r])
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.log = function (t) {
        Hh(t, 'log')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        ) {
          var o = e[r]
          n[r] = Math.log(o)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.log1p = function (t) {
        Hh(t, 'log1p')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        ) {
          var o = e[r]
          n[r] = Math.log1p(o)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.sqrt = function (t) {
        Hh(t, 'sqrt')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        ) {
          var o = e[r]
          n[r] = Math.sqrt(o)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.rsqrt = function (t) {
        Hh(t, 'rsqrt')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        ) {
          var o = e[r]
          n[r] = 1 / Math.sqrt(o)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.reciprocal = function (t) {
        Hh(t, 'reciprocal')
        for (
          var e = this.readSync(t.dataId),
            n = new Float32Array(e.length),
            r = 0;
          r < e.length;
          ++r
        )
          n[r] = 1 / e[r]
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.linear = function (t) {
        return t
      }),
      (o.prototype.relu = function (t) {
        Hh(t, 'relu')
        for (
          var e = Gn(t.shape, t.dtype),
            n = this.readSync(e.dataId),
            r = this.readSync(t.dataId),
            o = 0;
          o < r.length;
          ++o
        )
          n[o] = Math.max(0, r[o])
        return e
      }),
      (o.prototype.relu6 = function (t) {
        Hh(t, 'relu')
        for (
          var e = Gn(t.shape, t.dtype),
            n = this.readSync(e.dataId),
            r = this.readSync(t.dataId),
            o = 0;
          o < r.length;
          ++o
        )
          n[o] = Math.min(Math.max(0, r[o]), 6)
        return e
      }),
      (o.prototype.prelu = function (t, e) {
        return (
          Hh([t, e], 'prelu'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            return t < 0 ? e * t : t
          })
        )
      }),
      (o.prototype.elu = function (t) {
        Hh(t, 'elu')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        ) {
          var o = n[r]
          e[r] = o >= 0 ? o : Math.exp(o) - 1
        }
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.eluDer = function (t, e) {
        Hh([t, e], 'eluDer')
        for (
          var n = new Float32Array(e.size),
            r = this.readSync(e.dataId),
            o = this.readSync(t.dataId),
            a = 0;
          a < r.length;
          ++a
        ) {
          var i = r[a]
          n[a] = i >= 1 ? o[a] : o[a] * (i + 1)
        }
        return this.makeOutput(n, e.shape, 'float32')
      }),
      (o.prototype.selu = function (t) {
        Hh(t, 'selu')
        for (
          var e = bs,
            n = ws,
            r = new Float32Array(t.size),
            o = this.readSync(t.dataId),
            a = 0;
          a < o.length;
          ++a
        ) {
          var i = o[a]
          r[a] = i >= 0 ? n * i : e * (Math.exp(i) - 1)
        }
        return this.makeOutput(r, t.shape, 'float32')
      }),
      (o.prototype.clip = function (t, e, n) {
        Hh(t, 'clip')
        for (
          var r = new Float32Array(t.size), o = this.readSync(t.dataId), a = 0;
          a < o.length;
          ++a
        ) {
          var i = o[a]
          r[a] = i > n ? n : i < e ? e : i
        }
        return this.makeOutput(r, t.shape, 'float32')
      }),
      (o.prototype.abs = function (t) {
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.abs(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.complexAbs = function (t) {
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < t.size;
          ++r
        ) {
          var o = n[2 * r],
            a = n[2 * r + 1]
          e[r] = Math.hypot(o, a)
        }
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.int = function (t) {
        Hh(t, 'int')
        for (
          var e = new Int32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = n[r]
        return this.makeOutput(e, t.shape, 'int32')
      }),
      (o.prototype.sigmoid = function (t) {
        Hh(t, 'sigmoid')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = 1 / (1 + Math.exp(-n[r]))
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.softplus = function (t) {
        Hh(t, 'softplus')
        for (
          var e = Math.log(1.1920928955078125e-7) + 2,
            n = new Float32Array(t.size),
            r = this.readSync(t.dataId),
            o = 0;
          o < r.length;
          ++o
        ) {
          var a = r[o] > -e,
            i = r[o] < e,
            s = Math.exp(r[o]),
            u = void 0
          ;(u = i ? s : a ? r[o] : Math.log(1 + s)), (n[o] = u)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.sin = function (t) {
        Hh(t, 'sin')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.sin(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.cos = function (t) {
        Hh(t, 'cos')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.cos(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.tan = function (t) {
        Hh(t, 'tan')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.tan(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.asin = function (t) {
        Hh(t, 'asin')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.asin(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.acos = function (t) {
        Hh(t, 'acos')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.acos(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.atan = function (t) {
        Hh(t, 'atan')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.atan(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.atan2 = function (t, e) {
        return (
          Hh([t, e], 'atan2'),
          this.broadcastedBinaryOp(t, e, t.dtype, function (t, e) {
            return Math.atan2(t, e)
          })
        )
      }),
      (o.prototype.sinh = function (t) {
        Hh(t, 'sinh')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.sinh(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.cosh = function (t) {
        Hh(t, 'cosh')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.cosh(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.tanh = function (t) {
        Hh(t, 'tanh')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = D(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.asinh = function (t) {
        Hh(t, 'asinh')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.asinh(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.acosh = function (t) {
        Hh(t, 'acosh')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.acosh(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.atanh = function (t) {
        Hh(t, 'atanh')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        )
          e[r] = Math.atanh(n[r])
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.erf = function (t) {
        Hh(t, 'erf')
        for (
          var e = new Float32Array(t.size), n = this.readSync(t.dataId), r = 0;
          r < n.length;
          ++r
        ) {
          var o = Math.sign(n[r]),
            a = Math.abs(n[r]),
            i = 1 / (1 + 0.3275911 * a)
          e[r] =
            o *
            (1 -
              ((((1.061405429 * i - 1.453152027) * i + 1.421413741) * i -
                0.284496736) *
                i +
                0.254829592) *
                i *
                Math.exp(-a * a))
        }
        return this.makeOutput(e, t.shape, 'float32')
      }),
      (o.prototype.step = function (t, e) {
        void 0 === e && (e = 0), Hh(t, 'step')
        for (
          var n = new Float32Array(t.size), r = this.readSync(t.dataId), o = 0;
          o < r.length;
          ++o
        ) {
          var a = r[o]
          isNaN(a) ? (n[o] = NaN) : (n[o] = a > 0 ? 1 : e)
        }
        return this.makeOutput(n, t.shape, 'float32')
      }),
      (o.prototype.fusedConv2d = function (t) {
        var e = t.input,
          n = t.filter,
          r = t.convInfo,
          o = t.bias,
          a = t.activation,
          i = t.preluActivationWeights,
          s = this.conv2d(e, n, r)
        return o && (s = this.add(s, o)), a && (s = qh(this, s, a, i)), s
      }),
      (o.prototype.conv2d = function (t, e, n) {
        Hh([t, e], 'conv2d')
        for (
          var r = n.filterHeight,
            o = n.filterWidth,
            a = n.dilationHeight,
            i = n.dilationWidth,
            s = n.padInfo.left,
            u = n.padInfo.top,
            c = 'channelsLast' === n.dataFormat,
            l = dr(n.outShape, t.dtype),
            h = t.strides[0],
            f = c ? t.strides[1] : t.strides[2],
            d = c ? t.strides[2] : 1,
            p = c ? 1 : t.strides[1],
            v = l.strides[0],
            m = c ? l.strides[1] : l.strides[2],
            g = c ? l.strides[2] : 1,
            y = c ? 1 : l.strides[1],
            x = this.readSync(t.dataId),
            b = this.readSync(e.dataId),
            w = l.values,
            C = 0;
          C < n.batchSize;
          ++C
        )
          for (var E = C * h, R = C * v, I = 0; I < n.outHeight; ++I)
            for (
              var k = R + I * m, S = I * n.strideHeight - u, A = 0;
              A < r;
              A++
            ) {
              var D = S + A * a
              if (!(D < 0 || D >= n.inHeight))
                for (
                  var T = A * e.strides[0], N = E + D * f, F = 0;
                  F < n.outWidth;
                  ++F
                )
                  for (
                    var _ = k + F * g, O = F * n.strideWidth - s, M = 0;
                    M < o;
                    M++
                  ) {
                    var B = O + M * i
                    if (!(B < 0 || B >= n.inWidth))
                      for (
                        var P = N + B * d, L = T + M * e.strides[1], W = 0;
                        W < n.inChannels;
                        ++W
                      ) {
                        for (
                          var U = x[P + W * p], V = 0;
                          V < n.outChannels;
                          ++V
                        )
                          w[_ + V * y] += U * b[L + V]
                        L += n.outChannels
                      }
                  }
            }
        return l.toTensor()
      }),
      (o.prototype.conv3d = function (t, e, n) {
        for (
          var r = n.filterDepth,
            o = n.filterHeight,
            a = n.filterWidth,
            i = n.dilationDepth,
            s = n.dilationHeight,
            u = n.dilationWidth,
            c = n.padInfo.front,
            l = n.padInfo.left,
            h = n.padInfo.top,
            f = dr(n.outShape, t.dtype),
            d = this.readSync(t.dataId),
            p = this.readSync(e.dataId),
            v = f.values,
            m = 0;
          m < n.batchSize;
          ++m
        )
          for (
            var g = m * t.strides[0], y = m * f.strides[0], x = 0;
            x < n.outDepth;
            ++x
          )
            for (
              var b = y + x * f.strides[1], w = x * n.strideDepth - c, C = 0;
              C < r;
              C++
            ) {
              var E = w + C * i
              if (!(E < 0 || E >= n.inDepth))
                for (
                  var R = C * e.strides[0], I = g + E * t.strides[1], k = 0;
                  k < n.outHeight;
                  ++k
                )
                  for (
                    var S = b + k * f.strides[2],
                      A = k * n.strideHeight - h,
                      D = 0;
                    D < o;
                    D++
                  ) {
                    var T = A + D * s
                    if (!(T < 0 || T >= n.inHeight))
                      for (
                        var N = R + D * e.strides[1],
                          F = I + T * t.strides[2],
                          _ = 0;
                        _ < n.outWidth;
                        ++_
                      )
                        for (
                          var O = S + _ * n.outChannels,
                            M = _ * n.strideWidth - l,
                            B = 0;
                          B < a;
                          B++
                        ) {
                          var P = M + B * u
                          if (!(P < 0 || P >= n.inWidth))
                            for (
                              var L = N + B * e.strides[2],
                                W = F + P * n.inChannels,
                                U = L,
                                V = 0;
                              V < n.inChannels;
                              ++V
                            ) {
                              for (
                                var z = d[W + V], G = 0;
                                G < n.outChannels;
                                ++G
                              )
                                v[O + G] += z * p[U + G]
                              U += n.outChannels
                            }
                        }
                  }
            }
        return f.toTensor()
      }),
      (o.prototype.conv2dDerInput = function (t, e, n) {
        Hh([t, e], 'conv2dDerInput')
        for (
          var r = dr(n.inShape, 'float32'),
            o = r.values,
            a = this.readSync(t.dataId),
            i = this.readSync(e.dataId),
            s = e.strides,
            u = s[0],
            c = s[1],
            l = s[2],
            h = n.batchSize,
            f = n.filterHeight,
            d = n.filterWidth,
            p = n.inChannels,
            v = n.inHeight,
            m = n.inWidth,
            g = n.outChannels,
            y = n.outHeight,
            x = n.outWidth,
            b = n.strideHeight,
            w = n.strideWidth,
            C = n.dataFormat,
            E = f - 1 - n.padInfo.top,
            R = d - 1 - n.padInfo.left,
            I = 'channelsLast' === C,
            k = r.strides[0],
            S = I ? r.strides[1] : r.strides[2],
            A = I ? r.strides[2] : 1,
            D = I ? 1 : r.strides[1],
            T = t.strides[0],
            N = I ? t.strides[1] : t.strides[2],
            F = I ? t.strides[2] : 1,
            _ = I ? 1 : t.strides[1],
            O = 0;
          O < h;
          ++O
        )
          for (var M = 0; M < p; ++M)
            for (var B = 0; B < v; ++B)
              for (
                var P = B - E,
                  L = Math.max(0, Math.ceil(P / b)),
                  W = Math.min(y, (f + P) / b),
                  U = 0;
                U < m;
                ++U
              ) {
                for (
                  var V = U - R,
                    z = Math.max(0, Math.ceil(V / w)),
                    G = Math.min(x, (d + V) / w),
                    H = 0,
                    q = L;
                  q < W;
                  ++q
                )
                  for (var K = q * b - P, j = z; j < G; ++j)
                    for (
                      var X = T * O + N * q + F * j,
                        Y = u * (f - 1 - K) + c * (d - 1 - (j * w - V)) + l * M,
                        $ = 0;
                      $ < g;
                      ++$
                    ) {
                      H += a[X + _ * $] * i[Y + $]
                    }
                o[k * O + S * B + A * U + D * M] = H
              }
        return r.toTensor()
      }),
      (o.prototype.conv3dDerInput = function (t, e, n) {
        for (
          var r = dr(n.inShape, 'float32'),
            o = r.values,
            a = r.strides,
            i = a[0],
            s = a[1],
            u = a[2],
            c = a[3],
            l = this.readSync(t.dataId),
            h = t.strides,
            f = h[0],
            d = h[1],
            p = h[2],
            v = h[3],
            m = this.readSync(e.dataId),
            g = e.strides,
            y = g[0],
            x = g[1],
            b = g[2],
            w = g[3],
            C = n.batchSize,
            E = n.filterDepth,
            R = n.filterHeight,
            I = n.filterWidth,
            k = n.inChannels,
            S = n.inDepth,
            A = n.inHeight,
            D = n.inWidth,
            T = n.outChannels,
            N = n.outDepth,
            F = n.outHeight,
            _ = n.outWidth,
            O = n.strideDepth,
            M = n.strideHeight,
            B = n.strideWidth,
            P = E - 1 - n.padInfo.front,
            L = R - 1 - n.padInfo.top,
            W = I - 1 - n.padInfo.left,
            U = 0;
          U < C;
          ++U
        )
          for (var V = 0; V < k; ++V)
            for (var z = 0; z < S; ++z)
              for (
                var G = z - P,
                  H = Math.max(0, Math.ceil(G / O)),
                  q = Math.min(N, (E + G) / O),
                  K = 0;
                K < A;
                ++K
              )
                for (
                  var j = K - L,
                    X = Math.max(0, Math.ceil(j / M)),
                    Y = Math.min(F, (R + j) / M),
                    $ = 0;
                  $ < D;
                  ++$
                ) {
                  for (
                    var Q = $ - W,
                      J = Math.max(0, Math.ceil(Q / B)),
                      Z = Math.min(_, (I + Q) / B),
                      tt = 0,
                      et = H;
                    et < q;
                    ++et
                  )
                    for (var nt = et * O - G, rt = X; rt < Y; ++rt)
                      for (var ot = rt * M - j, at = J; at < Z; ++at)
                        for (
                          var it = f * U + d * et + p * rt + v * at,
                            st =
                              y * (E - 1 - nt) +
                              x * (R - 1 - ot) +
                              b * (I - 1 - (at * B - Q)) +
                              w * V,
                            ut = 0;
                          ut < T;
                          ++ut
                        ) {
                          tt += l[it + ut] * m[st + ut]
                        }
                  o[i * U + s * z + u * K + c * $ + V] = tt
                }
        return r.toTensor()
      }),
      (o.prototype.conv2dDerFilter = function (t, e, n) {
        Hh([t, e], 'conv2dDerFilter')
        for (
          var r = n.strideHeight,
            o = n.strideWidth,
            a = n.filterHeight,
            i = n.filterWidth,
            s = 'channelsLast' === n.dataFormat,
            u = dr(n.filterShape, 'float32'),
            c = n.padInfo.left,
            l = n.padInfo.top,
            h = this.bufferSync(t),
            f = this.bufferSync(e),
            d = 0;
          d < a;
          ++d
        )
          for (
            var p = Math.max(0, Math.ceil((l - d) / r)),
              v = Math.min(n.outHeight, (n.inHeight + l - d) / r),
              m = 0;
            m < i;
            ++m
          )
            for (
              var g = Math.max(0, Math.ceil((c - m) / o)),
                y = Math.min(n.outWidth, (n.inWidth + c - m) / o),
                x = 0;
              x < n.inChannels;
              ++x
            )
              for (var b = 0; b < n.outChannels; ++b) {
                for (var w = 0, C = 0; C < n.batchSize; ++C)
                  for (var E = p; E < v; ++E)
                    for (var R = d + E * r - l, I = g; I < y; ++I) {
                      var k = m + I * o - c
                      w += s
                        ? h.get(C, R, k, x) * f.get(C, E, I, b)
                        : h.get(C, x, R, k) * f.get(C, b, E, I)
                    }
                u.set(w, d, m, x, b)
              }
        return u.toTensor()
      }),
      (o.prototype.conv3dDerFilter = function (t, e, n) {
        for (
          var r = n.strideDepth,
            o = n.strideHeight,
            a = n.strideWidth,
            i = n.filterDepth,
            s = n.filterHeight,
            u = n.filterWidth,
            c = dr(n.filterShape, 'float32'),
            l = c.values,
            h = c.strides,
            f = h[0],
            d = h[1],
            p = h[2],
            v = h[3],
            m = this.readSync(e.dataId),
            g = e.strides,
            y = g[0],
            x = g[1],
            b = g[2],
            w = g[3],
            C = this.readSync(t.dataId),
            E = t.strides,
            R = E[0],
            I = E[1],
            k = E[2],
            S = E[3],
            A = n.padInfo.front,
            D = n.padInfo.left,
            T = n.padInfo.top,
            N = 0;
          N < i;
          ++N
        )
          for (
            var F = Math.max(0, Math.ceil((A - N) / r)),
              _ = Math.min(n.outDepth, (n.inDepth + A - N) / r),
              O = N * f,
              M = 0;
            M < s;
            ++M
          )
            for (
              var B = Math.max(0, Math.ceil((T - M) / o)),
                P = Math.min(n.outHeight, (n.inHeight + T - M) / o),
                L = M * d + O,
                W = 0;
              W < u;
              ++W
            )
              for (
                var U = Math.max(0, Math.ceil((D - W) / a)),
                  V = Math.min(n.outWidth, (n.inWidth + D - W) / a),
                  z = W * p + L,
                  G = 0;
                G < n.inChannels;
                ++G
              )
                for (var H = G * v + z, q = 0; q < n.outChannels; ++q) {
                  for (var K = 0, j = 0; j < n.batchSize; ++j)
                    for (var X = j * R, Y = j * y, $ = F; $ < _; ++$)
                      for (
                        var Q = (N + $ * r - A) * I + X, J = $ * x + Y, Z = B;
                        Z < P;
                        ++Z
                      )
                        for (
                          var tt = (M + Z * o - T) * k + Q,
                            et = Z * b + J,
                            nt = U;
                          nt < V;
                          ++nt
                        ) {
                          var rt = nt * w + et
                          K += C[(W + nt * a - D) * S + tt + G] * m[rt + q]
                        }
                  l[H + q] = K
                }
        return c.toTensor()
      }),
      (o.prototype.fusedDepthwiseConv2D = function (t) {
        var e = t.input,
          n = t.filter,
          r = t.convInfo,
          o = t.bias,
          a = t.activation,
          i = t.preluActivationWeights,
          s = this.depthwiseConv2D(e, n, r)
        return o && (s = this.add(s, o)), a && (s = qh(this, s, a, i)), s
      }),
      (o.prototype.depthwiseConv2D = function (t, e, n) {
        Hh([t, e], 'depthwiseConv2D')
        for (
          var r = n.filterHeight,
            o = n.filterWidth,
            a = n.dilationHeight,
            i = n.dilationWidth,
            s = n.padInfo.left,
            u = n.padInfo.top,
            c = n.outChannels / n.inChannels,
            l = dr(n.outShape, t.dtype),
            h = this.readSync(t.dataId),
            f = this.readSync(e.dataId),
            d = l.values,
            p = 0;
          p < n.batchSize;
          ++p
        )
          for (
            var v = p * t.strides[0], m = p * l.strides[0], g = 0;
            g < n.outHeight;
            ++g
          )
            for (
              var y = m + g * l.strides[1], x = g * n.strideHeight - s, b = 0;
              b < r;
              ++b
            ) {
              var w = x + b * a
              if (!(w < 0 || w >= n.inHeight))
                for (
                  var C = b * e.strides[0], E = v + w * t.strides[1], R = 0;
                  R < n.outWidth;
                  ++R
                )
                  for (
                    var I = y + R * l.strides[2],
                      k = R * n.strideWidth - u,
                      S = 0;
                    S < o;
                    ++S
                  ) {
                    var A = k + S * i
                    if (!(A < 0 || A >= n.inWidth))
                      for (
                        var D = C + S * e.strides[1],
                          T = E + A * n.inChannels,
                          N = I,
                          F = D,
                          _ = 0;
                        _ < n.inChannels;
                        ++_
                      ) {
                        for (var O = h[T + _], M = 0; M < c; ++M)
                          d[N + M] += O * f[F + M]
                        ;(N += c), (F += c)
                      }
                  }
            }
        return l.toTensor()
      }),
      (o.prototype.depthwiseConv2DDerInput = function (t, e, n) {
        Hh([t, e], 'depthwiseConv2DDerInput')
        for (
          var r = dr(n.inShape, 'float32'),
            o = r.values,
            a = r.strides,
            i = a[0],
            s = a[1],
            u = a[2],
            c = this.readSync(t.dataId),
            l = t.strides,
            h = l[0],
            f = l[1],
            d = l[2],
            p = this.readSync(e.dataId),
            v = e.strides,
            m = v[0],
            g = v[1],
            y = v[2],
            x = n.batchSize,
            b = n.filterHeight,
            w = n.filterWidth,
            C = n.inChannels,
            E = n.inHeight,
            R = n.inWidth,
            I = n.outChannels,
            k = n.outHeight,
            S = n.outWidth,
            A = n.strideHeight,
            D = n.strideWidth,
            T = b - 1 - n.padInfo.top,
            N = w - 1 - n.padInfo.left,
            F = I / C,
            _ = 0;
          _ < x;
          ++_
        )
          for (var O = 0; O < C; ++O)
            for (var M = 0; M < E; ++M)
              for (
                var B = M - T,
                  P = Math.max(0, Math.ceil(B / A)),
                  L = Math.min(k, (b + B) / A),
                  W = 0;
                W < R;
                ++W
              ) {
                for (
                  var U = W - N,
                    V = Math.max(0, Math.ceil(U / D)),
                    z = Math.min(S, (w + U) / D),
                    G = 0,
                    H = P;
                  H < L;
                  ++H
                )
                  for (var q = H * A - B, K = V; K < z; ++K)
                    for (
                      var j = h * _ + f * H + d * K,
                        X = m * (b - 1 - q) + g * (w - 1 - (K * D - U)) + y * O,
                        Y = 0;
                      Y < F;
                      ++Y
                    ) {
                      G += c[j + (O * F + Y)] * p[X + Y]
                    }
                o[i * _ + s * M + u * W + O] = G
              }
        return r.toTensor()
      }),
      (o.prototype.depthwiseConv2DDerFilter = function (t, e, n) {
        Hh([t, e], 'depthwiseConv2DDerFilter')
        for (
          var r = n.strideHeight,
            o = n.strideWidth,
            a = n.filterHeight,
            i = n.filterWidth,
            s = dr(n.filterShape, 'float32'),
            u = n.padInfo.left,
            c = n.padInfo.top,
            l = n.outChannels / n.inChannels,
            h = this.bufferSync(t),
            f = this.bufferSync(e),
            d = 0;
          d < a;
          ++d
        )
          for (
            var p = Math.max(0, Math.ceil((c - d) / r)),
              v = Math.min(n.outHeight, (n.inHeight + c - d) / r),
              m = 0;
            m < i;
            ++m
          )
            for (
              var g = Math.max(0, Math.ceil((u - m) / o)),
                y = Math.min(n.outWidth, (n.inWidth + u - m) / o),
                x = 0;
              x < n.outChannels;
              ++x
            ) {
              for (
                var b = Math.trunc(x / l), w = x % l, C = 0, E = 0;
                E < n.batchSize;
                ++E
              )
                for (var R = p; R < v; ++R)
                  for (var I = d + R * r - c, k = g; k < y; ++k) {
                    var S = m + k * o - u
                    C += h.get(E, I, S, b) * f.get(E, R, k, x)
                  }
              s.set(C, d, m, b, w)
            }
        return s.toTensor()
      }),
      (o.prototype.tile = function (t, e) {
        return Hh(t, 'tile'), ta(this.bufferSync(t), e)
      }),
      (o.prototype.pad = function (t, e, n) {
        Hh(t, 'pad')
        var r = e.map(function (e, n) {
            return e[0] + t.shape[n] + e[1]
          }),
          o = e.map(function (t) {
            return t[0]
          }),
          a = this.bufferSync(t),
          i = dr(r, t.dtype)
        0 !== n && i.values.fill(n)
        for (var s = 0; s < t.size; s++) {
          var u = a.indexToLoc(s),
            c = u.map(function (t, e) {
              return t + o[e]
            })
          i.set.apply(i, [a.get.apply(a, u)].concat(c))
        }
        return i.toTensor()
      }),
      (o.prototype.transpose = function (t, e) {
        Hh(t, 'transpose')
        for (var n = new Array(t.rank), r = 0; r < n.length; r++)
          n[r] = t.shape[e[r]]
        var o = this.readSync(t.dataId),
          a = dr(n, t.dtype),
          i = this.bufferSync(t)
        for (r = 0; r < t.size; ++r) {
          for (
            var s = i.indexToLoc(r), u = new Array(s.length), c = 0;
            c < u.length;
            c++
          )
            u[c] = s[e[c]]
          var l = a.locToIndex(u)
          a.values[l] = o[r]
        }
        return a.toTensor()
      }),
      (o.prototype.gather = function (t, e, n) {
        Hh([t, e], 'gather')
        var r = t.shape.slice(),
          o = this.readSync(e.dataId)
        r[n] = o.length
        for (
          var a = dr(r, t.dtype), i = this.bufferSync(t), s = 0;
          s < a.size;
          ++s
        ) {
          var u = a.indexToLoc(s),
            c = u.slice()
          c[n] = o[u[n]]
          var l = i.locToIndex(c)
          a.values[s] = i.values[l]
        }
        return a.toTensor()
      }),
      (o.prototype.batchToSpaceND = function (t, e, n) {
        Hh([t], 'batchToSpaceND')
        var r = e.reduce(function (t, e) {
            return t * e
          }),
          o = zr(t.shape, e, r),
          a = Gr(o.length, e.length),
          i = Hr(t.shape, e, r),
          s = qr(n, e.length),
          u = Kr(i, n, e.length)
        return t.reshape(o).transpose(a).reshape(i).slice(s, u)
      }),
      (o.prototype.spaceToBatchND = function (t, e, n) {
        Hh([t], 'spaceToBatchND')
        var r = e.reduce(function (t, e) {
            return t * e
          }),
          o = [[0, 0]]
        o.push.apply(o, n)
        for (var a = 1 + e.length; a < t.shape.length; ++a) o.push([0, 0])
        var i = t.pad(o),
          s = zr(i.shape, e, r, !1),
          u = Gr(s.length, e.length, !1),
          c = Hr(i.shape, e, r, !1)
        return i.reshape(s).transpose(u).reshape(c)
      }),
      (o.prototype.pool = function (t, e, n) {
        Hh(t, 'pool')
        for (
          var r = e.strideHeight,
            o = e.strideWidth,
            a = e.dilationHeight,
            i = e.dilationWidth,
            s = e.effectiveFilterHeight,
            u = e.effectiveFilterWidth,
            c = e.padInfo.top,
            l = e.padInfo.left,
            h =
              'max' === n ? Number.NEGATIVE_INFINITY : Number.POSITIVE_INFINITY,
            f = this.readSync(t.dataId),
            d = dr(e.outShape, t.dtype),
            p = d.values,
            v = e.outShape[1] * e.outShape[2] * e.outShape[3],
            m = e.outShape[2] * e.outShape[3],
            g = e.outShape[3],
            y = 0;
          y < e.batchSize;
          ++y
        )
          for (
            var x = y * v, b = y * t.strides[0], w = 0;
            w < e.inChannels;
            ++w
          )
            for (var C = 0; C < e.outHeight; ++C)
              for (
                var E = C * r - c,
                  R = Math.max(0, E),
                  I = Math.min(e.inHeight, s + E),
                  k = x + C * m,
                  S = 0;
                S < e.outWidth;
                ++S
              ) {
                for (
                  var A = S * o - l,
                    D = Math.max(0, A),
                    T = Math.min(e.inWidth, u + A),
                    N = h,
                    F = 0,
                    _ = 0,
                    O = R;
                  O < I;
                  O += a
                ) {
                  for (var M = b + O * t.strides[1], B = D; B < T; B += i) {
                    var P = f[M + B * t.strides[2] + w]
                    'max' === n && P > N
                      ? (N = P)
                      : 'avg' === n && ((F += P), _++)
                  }
                  if (isNaN(N)) break
                }
                p[k + S * g + w] = 'avg' === n ? F / _ : N
              }
        return d.toTensor()
      }),
      (o.prototype.maxPool = function (t, e) {
        return this.pool(t, e, 'max')
      }),
      (o.prototype.maxPoolPositions = function (t, e) {
        for (
          var n = dr(e.outShape, 'int32'),
            r = e.strideHeight,
            o = e.strideWidth,
            a = e.dilationHeight,
            i = e.dilationWidth,
            s = e.effectiveFilterHeight,
            u = e.effectiveFilterWidth,
            c = e.padInfo.top,
            l = e.padInfo.left,
            h = this.bufferSync(t),
            f = 0;
          f < e.batchSize;
          ++f
        )
          for (var d = 0; d < e.inChannels; ++d)
            for (var p = 0; p < e.outHeight; ++p) {
              for (var v = p * r - c, m = v; m < 0; ) m += a
              for (
                var g = Math.min(e.inHeight, s + v), y = 0;
                y < e.outWidth;
                ++y
              ) {
                for (var x = y * o - l, b = x; b < 0; ) b += i
                for (
                  var w = Math.min(e.inWidth, u + x),
                    C = Number.NEGATIVE_INFINITY,
                    E = -1,
                    R = m;
                  R < g;
                  R += a
                )
                  for (var I = R - v, k = b; k < w; k += i) {
                    var S = k - x,
                      A = h.get(f, R, k, d)
                    A > C && ((C = A), (E = I * u + S))
                  }
                n.set(E, f, p, y, d)
              }
            }
        return n.toTensor()
      }),
      (o.prototype.maxPoolBackprop = function (t, e, n, r) {
        Hh([e, n], 'maxPoolBackprop')
        for (
          var o = this.maxPoolPositions(e, r),
            a = r.strideHeight,
            i = r.strideWidth,
            s = r.dilationHeight,
            u = r.dilationWidth,
            c = r.effectiveFilterHeight,
            l = r.effectiveFilterWidth,
            h = l - 1 - r.padInfo.left,
            f = c - 1 - r.padInfo.top,
            d = dr(e.shape, 'float32'),
            p = this.bufferSync(o),
            v = this.bufferSync(t),
            m = 0;
          m < r.batchSize;
          ++m
        )
          for (var g = 0; g < r.inChannels; ++g)
            for (var y = 0; y < r.inHeight; ++y)
              for (var x = 0; x < r.inWidth; ++x) {
                for (var b = y - f, w = x - h, C = 0, E = 0; E < c; E += s) {
                  var R = (b + E) / a
                  if (!(R < 0 || R >= r.outHeight || Math.floor(R) !== R))
                    for (var I = 0; I < l; I += u) {
                      var k = (w + I) / i
                      if (!(k < 0 || k >= r.outWidth || Math.floor(k) !== k)) {
                        var S =
                          c * l - 1 - p.get(m, R, k, g) === E * l + I ? 1 : 0
                        if (0 !== S) C += v.get(m, R, k, g) * S
                      }
                    }
                }
                d.set(C, m, y, x, g)
              }
        return d.toTensor()
      }),
      (o.prototype.avgPoolBackprop = function (t, e, n) {
        Hh([t, e], 'avgPoolBackprop')
        for (
          var r = n.strideHeight,
            o = n.strideWidth,
            a = n.filterHeight,
            i = n.filterWidth,
            s = n.dilationHeight,
            u = n.dilationWidth,
            c = n.effectiveFilterHeight,
            l = n.effectiveFilterWidth,
            h = l - 1 - n.padInfo.left,
            f = c - 1 - n.padInfo.top,
            d = dr(e.shape, 'float32'),
            p = 1 / (a * i),
            v = this.bufferSync(t),
            m = 0;
          m < n.batchSize;
          ++m
        )
          for (var g = 0; g < n.inChannels; ++g)
            for (var y = 0; y < n.inHeight; ++y)
              for (var x = 0; x < n.inWidth; ++x) {
                for (var b = y - f, w = x - h, C = 0, E = 0; E < c; E += s) {
                  var R = (b + E) / r
                  if (!(R < 0 || R >= n.outHeight || Math.floor(R) !== R))
                    for (var I = 0; I < l; I += u) {
                      var k = (w + I) / o
                      if (!(k < 0 || k >= n.outWidth || Math.floor(k) !== k))
                        C += v.get(m, R, k, g)
                    }
                }
                d.set(C * p, m, y, x, g)
              }
        return d.toTensor()
      }),
      (o.prototype.pool3d = function (t, e, n) {
        Hh(t, 'pool3d')
        for (
          var r = e.strideDepth,
            o = e.strideHeight,
            a = e.strideWidth,
            i = e.dilationDepth,
            s = e.dilationHeight,
            u = e.dilationWidth,
            c = e.effectiveFilterDepth,
            l = e.effectiveFilterHeight,
            h = e.effectiveFilterWidth,
            f = e.padInfo.front,
            d = e.padInfo.top,
            p = e.padInfo.left,
            v =
              'max' === n ? Number.NEGATIVE_INFINITY : Number.POSITIVE_INFINITY,
            m = this.readSync(t.dataId),
            g = dr(e.outShape, t.dtype),
            y = g.values,
            x = e.outShape[1] * e.outShape[2] * e.outShape[3] * e.outShape[4],
            b = e.outShape[2] * e.outShape[3] * e.outShape[4],
            w = e.outShape[3] * e.outShape[4],
            C = e.outShape[4],
            E = 0;
          E < e.batchSize;
          ++E
        )
          for (
            var R = E * x, I = E * t.strides[0], k = 0;
            k < e.inChannels;
            ++k
          )
            for (var S = 0; S < e.outDepth; ++S) {
              for (var A = S * r - f, D = A; D < 0; ) D += i
              for (
                var T = Math.min(e.inDepth, c + A), N = R + S * b, F = 0;
                F < e.outHeight;
                ++F
              ) {
                for (var _ = F * o - d, O = _; O < 0; ) O += s
                for (
                  var M = Math.min(e.inHeight, l + _), B = N + F * w, P = 0;
                  P < e.outWidth;
                  ++P
                ) {
                  for (var L = P * a - p, W = L; W < 0; ) W += u
                  for (
                    var U = Math.min(e.inWidth, h + L),
                      V = B + P * C,
                      z = v,
                      G = 0,
                      H = 0,
                      q = D;
                    q < T;
                    q += i
                  ) {
                    for (var K = I + q * t.strides[1], j = O; j < M; j += s) {
                      for (var X = K + j * t.strides[2], Y = W; Y < U; Y += u) {
                        var $ = m[X + Y * t.strides[3] + k]
                        if (
                          ('max' === n && $ > z
                            ? (z = $)
                            : 'avg' === n && ((G += $), H++),
                          isNaN(z))
                        )
                          break
                      }
                      if (isNaN(z)) break
                    }
                    if (isNaN(z)) break
                  }
                  y[V + k] = 'avg' === n ? G / H : z
                }
              }
            }
        return g.toTensor()
      }),
      (o.prototype.avgPool3d = function (t, e) {
        return Hh(t, 'avgPool3d'), this.pool3d(t, e, 'avg').toFloat()
      }),
      (o.prototype.avgPool3dBackprop = function (t, e, n) {
        Hh([t, e], 'avgPool3dBackprop')
        for (
          var r = n.strideDepth,
            o = n.strideHeight,
            a = n.strideWidth,
            i = n.filterDepth,
            s = n.filterHeight,
            u = n.filterWidth,
            c = n.dilationDepth,
            l = n.dilationHeight,
            h = n.dilationWidth,
            f = n.effectiveFilterDepth,
            d = n.effectiveFilterHeight,
            p = n.effectiveFilterWidth,
            v = f - 1 - n.padInfo.front,
            m = p - 1 - n.padInfo.left,
            g = d - 1 - n.padInfo.top,
            y = dr(e.shape, 'float32'),
            x = 1 / (i * s * u),
            b = this.bufferSync(t),
            w = 0;
          w < n.batchSize;
          ++w
        )
          for (var C = 0; C < n.inChannels; ++C)
            for (var E = 0; E < n.inDepth; ++E)
              for (var R = 0; R < n.inHeight; ++R)
                for (var I = 0; I < n.inWidth; ++I) {
                  for (
                    var k = E - v, S = R - g, A = I - m, D = 0, T = 0;
                    T < f;
                    T += c
                  ) {
                    var N = (k + T) / r
                    if (!(N < 0 || N >= n.outDepth || Math.floor(N) !== N))
                      for (var F = 0; F < d; F += l) {
                        var _ = (S + F) / o
                        if (!(_ < 0 || _ >= n.outHeight || Math.floor(_) !== _))
                          for (var O = 0; O < p; O += h) {
                            var M = (A + O) / a
                            if (
                              !(M < 0 || M >= n.outWidth || Math.floor(M) !== M)
                            )
                              D += b.get(w, N, _, M, C)
                          }
                      }
                  }
                  y.set(D * x, w, E, R, I, C)
                }
        return y.toTensor()
      }),
      (o.prototype.maxPool3d = function (t, e) {
        return Hh(t, 'maxPool3d'), this.pool3d(t, e, 'max').toFloat()
      }),
      (o.prototype.maxPool3dPositions = function (t, e) {
        for (
          var n = dr(e.outShape, 'int32'),
            r = e.strideDepth,
            o = e.strideHeight,
            a = e.strideWidth,
            i = e.dilationDepth,
            s = e.dilationHeight,
            u = e.dilationWidth,
            c = e.effectiveFilterDepth,
            l = e.effectiveFilterHeight,
            h = e.effectiveFilterWidth,
            f = e.padInfo.front,
            d = e.padInfo.top,
            p = e.padInfo.left,
            v = this.bufferSync(t),
            m = 0;
          m < e.batchSize;
          ++m
        )
          for (var g = 0; g < e.inChannels; ++g)
            for (var y = 0; y < e.outDepth; ++y) {
              for (var x = y * r - f, b = x; b < 0; ) b += i
              for (
                var w = Math.min(e.inDepth, c + x), C = 0;
                C < e.outHeight;
                ++C
              ) {
                for (var E = C * o - d, R = E; R < 0; ) R += s
                for (
                  var I = Math.min(e.inHeight, l + E), k = 0;
                  k < e.outWidth;
                  ++k
                ) {
                  for (var S = k * a - p, A = S; A < 0; ) A += u
                  for (
                    var D = Math.min(e.inWidth, h + S),
                      T = Number.NEGATIVE_INFINITY,
                      N = -1,
                      F = b;
                    F < w;
                    F += i
                  )
                    for (var _ = F - x, O = R; O < I; O += s)
                      for (var M = O - E, B = A; B < D; B += u) {
                        var P = B - S,
                          L = v.get(m, F, O, B, g)
                        L >= T && ((T = L), (N = _ * l * h + M * l + P))
                      }
                  n.set(N, m, y, C, k, g)
                }
              }
            }
        return n.toTensor()
      }),
      (o.prototype.maxPool3dBackprop = function (t, e, n, r) {
        Hh([e, n], 'maxPool3dBackprop')
        for (
          var o = this.maxPool3dPositions(e, r),
            a = r.strideDepth,
            i = r.strideHeight,
            s = r.strideWidth,
            u = r.dilationDepth,
            c = r.dilationHeight,
            l = r.dilationWidth,
            h = r.effectiveFilterDepth,
            f = r.effectiveFilterHeight,
            d = r.effectiveFilterWidth,
            p = h - 1 - r.padInfo.front,
            v = d - 1 - r.padInfo.left,
            m = f - 1 - r.padInfo.top,
            g = dr(e.shape, 'float32'),
            y = this.bufferSync(o),
            x = this.bufferSync(t),
            b = 0;
          b < r.batchSize;
          ++b
        )
          for (var w = 0; w < r.inChannels; ++w)
            for (var C = 0; C < r.inDepth; ++C)
              for (var E = 0; E < r.inHeight; ++E)
                for (var R = 0; R < r.inWidth; ++R) {
                  for (
                    var I = C - p, k = E - m, S = R - v, A = 0, D = 0;
                    D < h;
                    D += u
                  ) {
                    var T = (I + D) / a
                    if (!(T < 0 || T >= r.outDepth || Math.floor(T) !== T))
                      for (var N = 0; N < f; N += c) {
                        var F = (k + N) / i
                        if (!(F < 0 || F >= r.outHeight || Math.floor(F) !== F))
                          for (var _ = 0; _ < d; _ += l) {
                            var O = (S + _) / s
                            if (
                              !(O < 0 || O >= r.outWidth || Math.floor(O) !== O)
                            ) {
                              var M =
                                h * f * d - 1 - y.get(b, T, F, O, w) ===
                                D * f * d + N * d + _
                                  ? 1
                                  : 0
                              if (0 !== M) A += x.get(b, T, F, O, w) * M
                            }
                          }
                      }
                  }
                  g.set(A, b, C, E, R, w)
                }
        return g.toTensor()
      }),
      (o.prototype.cast = function (t, e) {
        return Po(t, e, this)
      }),
      (o.prototype.reshape = function (t, e) {
        return Lo(t, e)
      }),
      (o.prototype.avgPool = function (t, e) {
        return Hh(t, 'avgPool'), this.pool(t, e, 'avg').toFloat()
      }),
      (o.prototype.resizeBilinear = function (t, e, n, r) {
        Hh(t, 'resizeBilinear')
        for (
          var o = t.shape,
            a = o[0],
            i = o[1],
            s = o[2],
            u = o[3],
            c = this.readSync(t.dataId),
            l = new Float32Array(k([a, e, n, u])),
            h = [r && e > 1 ? i - 1 : i, r && n > 1 ? s - 1 : s],
            f = [r && e > 1 ? e - 1 : e, r && n > 1 ? n - 1 : n],
            d = 0,
            p = h[0] / f[0],
            v = h[1] / f[1],
            m = 0;
          m < a;
          m++
        )
          for (var g = 0; g < e; g++)
            for (
              var y = p * g,
                x = Math.floor(y),
                b = y - x,
                w = Math.min(i - 1, Math.ceil(y)),
                C = m * t.strides[0] + x * t.strides[1],
                E = m * t.strides[0] + w * t.strides[1],
                R = 0;
              R < n;
              R++
            )
              for (
                var I = v * R,
                  S = Math.floor(I),
                  A = I - S,
                  D = Math.min(s - 1, Math.ceil(I)),
                  T = C + S * t.strides[2],
                  N = E + S * t.strides[2],
                  F = C + D * t.strides[2],
                  _ = E + D * t.strides[2],
                  O = 0;
                O < u;
                O++
              ) {
                var M = c[T + O],
                  B = c[N + O],
                  P = M + (c[F + O] - M) * A,
                  L = P + (B + (c[_ + O] - B) * A - P) * b
                l[d++] = L
              }
        return Fn(l, [a, e, n, u])
      }),
      (o.prototype.resizeBilinearBackprop = function (t, e, n) {
        Hh([t, e], 'resizeBilinearBackprop')
        for (
          var r = e.shape,
            o = r[0],
            a = r[1],
            i = r[2],
            s = r[3],
            u = t.shape,
            c = u[1],
            l = u[2],
            h = new Float32Array(o * a * i * s),
            f = [n && c > 1 ? a - 1 : a, n && l > 1 ? i - 1 : i],
            d = [n && c > 1 ? c - 1 : c, n && l > 1 ? l - 1 : l],
            p = f[0] / d[0],
            v = f[1] / d[1],
            m = this.readSync(t.dataId),
            g = 0,
            y = 0;
          y < o;
          y++
        )
          for (var x = y * e.strides[0], b = 0; b < c; b++)
            for (
              var w = b * p,
                C = Math.floor(w),
                E = Math.min(Math.ceil(w), a - 1),
                R = x + C * e.strides[1],
                I = x + E * e.strides[1],
                k = w - C,
                S = 1 - k,
                A = 0;
              A < l;
              A++
            )
              for (
                var D = A * v,
                  T = Math.floor(D),
                  N = Math.min(Math.ceil(D), i - 1),
                  F = D - T,
                  _ = 1 - F,
                  O = R + T * e.strides[2],
                  M = R + N * e.strides[2],
                  B = I + T * e.strides[2],
                  P = I + N * e.strides[2],
                  L = S * _,
                  W = S * F,
                  U = k * _,
                  V = k * F,
                  z = 0;
                z < s;
                z++
              ) {
                var G = m[g++]
                ;(h[O + z] += G * L),
                  (h[M + z] += G * W),
                  (h[B + z] += G * U),
                  (h[P + z] += G * V)
              }
        return Ln(h, [o, i, a, s], e.dtype)
      }),
      (o.prototype.resizeNearestNeighbor = function (t, e, n, r) {
        Hh(t, 'resizeNearestNeighbor')
        for (
          var o = t.shape,
            a = o[0],
            i = o[1],
            s = o[2],
            u = o[3],
            c = this.readSync(t.dataId),
            l = new Float32Array(a * e * n * u),
            h = [r && e > 1 ? i - 1 : i, r && n > 1 ? s - 1 : s],
            f = [r && e > 1 ? e - 1 : e, r && n > 1 ? n - 1 : n],
            d = h[0] / f[0],
            p = h[1] / f[1],
            v = 0,
            m = 0;
          m < a;
          m++
        )
          for (var g = m * t.strides[0], y = 0; y < e; y++)
            for (
              var x = d * y,
                b =
                  g +
                  Math.min(i - 1, r ? Math.round(x) : Math.floor(x)) *
                    t.strides[1],
                w = 0;
              w < n;
              w++
            )
              for (
                var C = p * w,
                  E =
                    b +
                    Math.min(s - 1, r ? Math.round(C) : Math.floor(C)) *
                      t.strides[2],
                  R = 0;
                R < u;
                R++
              ) {
                var I = c[E + R]
                l[v++] = I
              }
        return Fn(l, [a, e, n, u], t.dtype)
      }),
      (o.prototype.resizeNearestNeighborBackprop = function (t, e, n) {
        Hh([t, e], 'resizeNearestNeighborBackprop')
        for (
          var r = e.shape,
            o = r[0],
            a = r[1],
            i = r[2],
            s = r[3],
            u = t.shape,
            c = u[1],
            l = u[2],
            h = new Float32Array(o * a * i * s),
            f = this.readSync(t.dataId),
            d = [n && c > 1 ? a - 1 : a, n && l > 1 ? i - 1 : i],
            p = [n && c > 1 ? c - 1 : c, n && l > 1 ? l - 1 : l],
            v = d[0] / p[0],
            m = d[1] / p[1],
            g = 1 / v,
            y = 1 / m,
            x = 2 * Math.ceil(g) + 2,
            b = 2 * Math.ceil(y) + 2,
            w = 0;
          w < o;
          w++
        )
          for (var C = w * e.strides[0], E = 0; E < a; E++)
            for (
              var R = C + E * e.strides[1],
                I = Math.floor(E * g),
                k = Math.floor(I - x / 2),
                S = 0;
              S < i;
              S++
            )
              for (
                var A = R + S * e.strides[2],
                  D = Math.floor(S * y),
                  T = Math.floor(D - b / 2),
                  N = 0;
                N < s;
                N++
              ) {
                for (var F = 0, _ = 0; _ < x; _++) {
                  var O = _ + k
                  if (!(O < 0 || O >= c)) {
                    var M = C + O * t.strides[1],
                      B = O * v
                    if (
                      E === Math.min(a - 1, n ? Math.round(B) : Math.floor(B))
                    )
                      for (var P = 0; P < b; P++) {
                        var L = P + T
                        if (!(L < 0 || L >= l)) {
                          var W = M + L * t.strides[2],
                            U = L * m
                          S ===
                            Math.min(
                              i - 1,
                              n ? Math.round(U) : Math.floor(U)
                            ) && (F += f[W + N])
                        }
                      }
                  }
                }
                h[A + N] = F
              }
        return Ln(h, e.shape, e.dtype)
      }),
      (o.prototype.batchNormalization = function (t, e, n, r, o, a) {
        Hh([t, e, n, o, a], 'batchNorm')
        for (
          var i = this.readSync(t.dataId),
            s = this.readSync(e.dataId),
            u = this.readSync(n.dataId),
            c = o ? this.readSync(o.dataId) : new Float32Array([1]),
            l = a ? this.readSync(a.dataId) : new Float32Array([0]),
            h = new Float32Array(i.length),
            f = l.length,
            d = c.length,
            p = u.length,
            v = s.length,
            m = 0,
            g = 0,
            y = 0,
            x = 0,
            b = 0;
          b < i.length;
          ++b
        )
          (h[b] = l[m++] + ((i[b] - s[g++]) * c[y++]) / Math.sqrt(u[x++] + r)),
            m >= f && (m = 0),
            g >= v && (g = 0),
            y >= d && (y = 0),
            x >= p && (x = 0)
        return Ln(h, t.shape)
      }),
      (o.prototype.localResponseNormalization4D = function (t, e, n, r, o) {
        Hh(t, 'localResponseNormalization4D')
        var a = t.shape[3],
          i = a - 1,
          s = this.readSync(t.dataId),
          u = t.size,
          c = new Float32Array(u)
        function l(t) {
          for (
            var n = t % a,
              r = t - n + Math.max(0, n - e),
              o = t - n + Math.min(n + e, i),
              u = 0;
            r <= o;
            r++
          ) {
            var c = s[r]
            u += c * c
          }
          return u
        }
        for (var h = 0; h < u; h++) {
          var f = l(h),
            d = s[h] * Math.pow(n + r * f, -o)
          c[h] = d
        }
        return Ln(c, t.shape)
      }),
      (o.prototype.LRNGrad = function (t, e, n, r, o, a, i) {
        Hh(t, 'LRNGrad')
        for (
          var s = t.shape[3],
            u = this.readSync(t.dataId),
            c = this.readSync(e.dataId),
            l = this.readSync(n.dataId),
            h = new Float32Array(t.size),
            f = t.size,
            d = 0;
          d < f;
          d++
        ) {
          for (
            var p = d % s,
              v = d - p + Math.max(0, p - r),
              m = d - p + Math.min(s, p + r + 1),
              g = 0,
              y = v;
            y < m;
            y++
          )
            g += Math.pow(c[y], 2)
          g = a * g + o
          for (y = v; y < m; y++) {
            var x = (-2 * a * i * c[y] * l[d]) / g
            d === y && (x += Math.pow(g, -i)), (x *= u[d]), (h[y] += x)
          }
        }
        return Ln(h, t.shape)
      }),
      (o.prototype.multinomial = function (t, e, n, r) {
        Hh(t, 'multinomial')
        for (
          var o = e ? t : go(t),
            a = o.shape[0],
            i = o.shape[1],
            s = Gn([a, n], 'int32'),
            u = this.readSync(s.dataId),
            c = this.readSync(o.dataId),
            l = 0;
          l < a;
          ++l
        ) {
          var h = l * i,
            f = new Float32Array(i - 1)
          f[0] = c[h]
          for (var d = 1; d < f.length; ++d) f[d] = f[d - 1] + c[h + d]
          for (var p = cr(r.toString()), v = l * n, m = 0; m < n; ++m) {
            var g = p()
            u[v + m] = f.length
            for (var y = 0; y < f.length; y++)
              if (g < f[y]) {
                u[v + m] = y
                break
              }
          }
        }
        return s
      }),
      (o.prototype.oneHot = function (t, e, n, r) {
        Hh(t, 'oneHot')
        var o = new Float32Array(t.size * e)
        o.fill(r)
        for (var a = this.readSync(t.dataId), i = 0; i < t.size; ++i)
          a[i] >= 0 && a[i] < e && (o[i * e + a[i]] = n)
        return Bn(o, [t.size, e], 'int32')
      }),
      (o.prototype.nonMaxSuppression = function (t, e, n, r, o) {
        return (
          Hh(t, 'nonMaxSuppression'),
          jo(this.readSync(t.dataId), this.readSync(e.dataId), n, r, o)
        )
      }),
      (o.prototype.fft = function (t) {
        return this.fftBatch(t, !1)
      }),
      (o.prototype.ifft = function (t) {
        return this.fftBatch(t, !0)
      }),
      (o.prototype.fftBatch = function (t, e) {
        for (
          var n = t.shape[0],
            r = t.shape[1],
            o = dr(t.shape, 'float32'),
            a = dr(t.shape, 'float32'),
            i = Tn(t).as2D(n, r),
            s = Nn(t).as2D(n, r),
            u = 0;
          u < n;
          u++
        )
          for (
            var c = i.slice([u, 0], [1, r]),
              l = s.slice([u, 0], [1, r]),
              h = Dn(c, l),
              f = this.readSync(this.fftImpl(h, e).dataId),
              d = 0;
            d < r;
            d++
          ) {
            var p = zo(f, d)
            ;(o.values[u * r + d] = p.real), (a.values[u * r + d] = p.imag)
          }
        return Dn(o.toTensor(), a.toTensor()).as2D(n, r)
      }),
      (o.prototype.fftImpl = function (t, e) {
        var n = t.as1D(),
          r = n.size
        if (this.isExponentOf2(r)) {
          var o = this.fftRadix2(n, r, e).as2D(t.shape[0], t.shape[1])
          return e && (o = Dn(Tn(o).div(On(r)), Nn(o).div(On(r)))), o
        }
        var a = this.readSync(t.dataId),
          i = (function (t) {
            for (
              var e = new Float32Array(t.length / 2),
                n = new Float32Array(t.length / 2),
                r = 0;
              r < t.length;
              r += 2
            )
              (e[r / 2] = t[r]), (n[r / 2] = t[r + 1])
            return { real: e, imag: n }
          })(this.fourierTransformByMatmul(a, r, e))
        return Dn(i.real, i.imag).as2D(t.shape[0], t.shape[1])
      }),
      (o.prototype.isExponentOf2 = function (t) {
        return 0 == (t & (t - 1))
      }),
      (o.prototype.fftRadix2 = function (t, e, n) {
        if (1 === e) return t
        var r = this.readSync(t.dataId),
          o = e / 2,
          a = (function (t) {
            for (
              var e = Math.ceil(t.length / 4),
                n = new Float32Array(e),
                r = new Float32Array(e),
                o = 0;
              o < t.length;
              o += 4
            )
              (n[Math.floor(o / 4)] = t[o]), (r[Math.floor(o / 4)] = t[o + 1])
            return { real: n, imag: r }
          })(r),
          i = Dn(a.real, a.imag).as1D(),
          s = (function (t) {
            for (
              var e = Math.floor(t.length / 4),
                n = new Float32Array(e),
                r = new Float32Array(e),
                o = 2;
              o < t.length;
              o += 4
            )
              (n[Math.floor(o / 4)] = t[o]), (r[Math.floor(o / 4)] = t[o + 1])
            return { real: n, imag: r }
          })(r),
          u = Dn(s.real, s.imag).as1D()
        ;(i = this.fftRadix2(i, o, n)), (u = this.fftRadix2(u, o, n))
        var c = (function (t, e) {
            for (
              var n = new Float32Array(t / 2),
                r = new Float32Array(t / 2),
                o = 0;
              o < Math.ceil(t / 2);
              o++
            ) {
              var a = (e ? 2 : -2) * Math.PI * (o / t)
              ;(n[o] = Math.cos(a)), (r[o] = Math.sin(a))
            }
            return { real: n, imag: r }
          })(e, n),
          l = Dn(c.real, c.imag).mul(u),
          h = i.add(l),
          f = i.sub(l),
          d = Tn(h).concat(Tn(f)),
          p = Nn(h).concat(Nn(f))
        return Dn(d, p).as1D()
      }),
      (o.prototype.fourierTransformByMatmul = function (t, e, n) {
        for (var r = new Float32Array(2 * e), o = 0; o < e; o++) {
          for (var a = 0, i = 0, s = 0; s < e; s++) {
            var u = Ho(o * s, e, n),
              c = zo(t, s)
            ;(a += c.real * u.real - c.imag * u.imag),
              (i += c.real * u.imag + c.imag * u.real)
          }
          n && ((a /= e), (i /= e)), Go(r, a, i, o)
        }
        return r
      }),
      (o.prototype.depthToSpace = function (t, e, n) {
        C('NHWC' === n, function () {
          return (
            'Only NHWC dataFormat supported on CPU for depthToSpace. Got ' + n
          )
        }),
          C(e > 1, function () {
            return 'blockSize should be > 1 for depthToSpace, but was: ' + e
          })
        for (
          var r = t.shape[0],
            o = t.shape[1],
            a = t.shape[2],
            i = t.shape[3],
            s = o * e,
            u = a * e,
            c = i / (e * e),
            l = this.readSync(t.dataId),
            h = new Float32Array(r * s * u * c),
            f = 0,
            d = 0;
          d < r;
          ++d
        )
          for (var p = 0; p < s; ++p)
            for (var v = Math.floor(p / e), m = p % e, g = 0; g < u; ++g)
              for (
                var y = Math.floor(g / e), x = (m * e + (g % e)) * c, b = 0;
                b < c;
                ++b
              ) {
                var w = b + x + i * (y + a * (v + o * d))
                h[f++] = l[w]
              }
        return Ln(h, [r, s, u, c])
      }),
      (o.prototype.broadcastedBinaryOp = function (t, e, n, r) {
        var o = Ro(t.shape, e.shape),
          a = dr(o, n),
          i = this.readSync(t.dataId),
          s = this.readSync(e.dataId),
          u = Co(t.shape, o),
          c = Co(e.shape, o),
          l = a.values
        if (u.length + c.length === 0)
          for (var h = 0; h < l.length; ++h)
            l[h] = r(i[h % i.length], s[h % s.length])
        else {
          var f = this.bufferSync(t),
            d = this.bufferSync(e),
            p = function (n) {
              var o = a.indexToLoc(n),
                h = o.slice(-t.rank)
              u.forEach(function (t) {
                return (h[t] = 0)
              })
              var p = f.locToIndex(h),
                v = o.slice(-e.rank)
              c.forEach(function (t) {
                return (v[t] = 0)
              })
              var m = d.locToIndex(v)
              l[n] = r(i[p], s[m])
            }
          for (h = 0; h < l.length; ++h) p(h)
        }
        return a.toTensor()
      }),
      (o.prototype.broadcastedBinaryComplexOp = function (t, e, n) {
        var r = Ro(t.shape, e.shape),
          o = dr(r, 'float32'),
          a = dr(r, 'float32'),
          i = this.readSync(t.dataId),
          s = this.readSync(e.dataId),
          u = Co(t.shape, r),
          c = Co(e.shape, r),
          l = o.values,
          h = a.values
        if (u.length + c.length === 0)
          for (var f = 0; f < l.length; f++) {
            var d = f % i.length,
              p = f % s.length,
              v = n(i[2 * d], i[2 * d + 1], s[2 * p], s[2 * p + 1])
            ;(l[f] = v.real), (h[f] = v.imag)
          }
        else {
          var m = this.bufferSync(this.data.get(t.dataId).complexTensors.real),
            g = this.bufferSync(this.data.get(e.dataId).complexTensors.real),
            y = function (r) {
              var a = o.indexToLoc(r),
                f = a.slice(-t.rank)
              u.forEach(function (t) {
                return (f[t] = 0)
              })
              var d = m.locToIndex(f),
                p = a.slice(-e.rank)
              c.forEach(function (t) {
                return (p[t] = 0)
              })
              var v = g.locToIndex(p),
                y = n(i[2 * d], i[2 * d + 1], s[2 * v], s[2 * v + 1])
              ;(l[r] = y.real), (h[r] = y.imag)
            }
          for (f = 0; f < l.length; f++) y(f)
        }
        return this.complex(o.toTensor(), a.toTensor())
      }),
      (o.prototype.split = function (t, e, n) {
        return Zo(t, e, n)
      }),
      (o.prototype.dispose = function () {}),
      (o.prototype.floatPrecision = function () {
        return 32
      }),
      (o.prototype.epsilon = function () {
        return 1e-7
      }),
      (o.prototype.cropAndResize = function (t, e, n, r, o, a) {
        for (
          var i = t.shape,
            s = i[0],
            u = i[1],
            c = i[2],
            l = i[3],
            h = e.shape[0],
            f = r[0],
            d = r[1],
            p = dr([h, f, d, l], 'float32'),
            v = this.readSync(e.dataId),
            m = this.readSync(n.dataId),
            g = this.readSync(t.dataId),
            y = t.strides,
            x = p.strides,
            b = 0;
          b < h;
          b++
        ) {
          var w = 4 * b,
            C = v[w],
            E = v[w + 1],
            R = v[w + 2],
            I = v[w + 3],
            k = m[b]
          if (!(k >= s))
            for (
              var S = f > 1 ? ((R - C) * (u - 1)) / (f - 1) : 0,
                A = d > 1 ? ((I - E) * (c - 1)) / (d - 1) : 0,
                D = 0;
              D < f;
              D++
            ) {
              var T = f > 1 ? C * (u - 1) + D * S : 0.5 * (C + R) * (u - 1)
              if (T < 0 || T > u - 1)
                for (var N = 0; N < d; N++)
                  for (var F = 0; F < l; F++) {
                    var _ = F + N * x[2] + D * x[1] + b * x[0]
                    p.values[_] = a
                  }
              else if ('bilinear' === o) {
                var O = Math.floor(T),
                  M = Math.ceil(T),
                  B = T - O
                for (N = 0; N < d; N++) {
                  if (
                    (q =
                      d > 1 ? E * (c - 1) + N * A : 0.5 * (E + I) * (c - 1)) <
                      0 ||
                    q > c - 1
                  )
                    for (F = 0; F < l; F++) {
                      _ = F + N * x[2] + D * x[1] + b * x[0]
                      p.values[_] = a
                    }
                  else {
                    var P = Math.floor(q),
                      L = Math.ceil(q),
                      W = q - P
                    for (F = 0; F < l; F++) {
                      var U = g[(_ = F + P * y[2] + O * y[1] + k * y[0])],
                        V = g[(_ = F + L * y[2] + O * y[1] + k * y[0])],
                        z = g[(_ = F + P * y[2] + M * y[1] + k * y[0])],
                        G = U + (V - U) * W,
                        H =
                          z +
                          (g[(_ = F + L * y[2] + M * y[1] + k * y[0])] - z) * W
                      ;(_ = F + N * x[2] + D * x[1] + b * x[0]),
                        (p.values[_] = G + (H - G) * B)
                    }
                  }
                }
              } else
                for (N = 0; N < d; ++N) {
                  var q
                  if (
                    (q =
                      d > 1 ? E * (c - 1) + N * A : 0.5 * (E + I) * (c - 1)) <
                      0 ||
                    q > c - 1
                  )
                    for (F = 0; F < l; F++) {
                      _ = F + N * x[2] + D * x[1] + b * x[0]
                      p.values[_] = a
                    }
                  else {
                    var K = Math.round(q),
                      j = Math.round(T)
                    for (F = 0; F < l; F++) {
                      var X = F + K * y[2] + j * y[1] + k * y[0],
                        Y = F + N * x[2] + D * x[1] + b * x[0]
                      p.values[Y] = g[X]
                    }
                  }
                }
            }
        }
        return p.toTensor()
      }),
      (o.prototype.sparseToDense = function (t, e, n, r) {
        var o = Zr(0, t, n),
          a = o.sliceRank,
          i = o.numUpdates,
          s = o.sliceSize,
          u = o.strides,
          c = o.outputSize
        return this.scatter(t, e, n, c, s, i, a, u, r, !1)
      }),
      (o.prototype.gatherND = function (t, e) {
        var n = e.shape,
          r = n[n.length - 1],
          o = jr(t, e),
          a = o[0],
          i = o[1],
          s = o[2],
          u = o[3]
        if (0 === i) return Fn([], a, t.dtype)
        for (
          var c = new gt([i, s], t.dtype),
            l = this.readSync(e.dataId),
            h = this.readSync(t.dataId),
            f = 0;
          f < i;
          f++
        ) {
          for (var d = [], p = 0, v = 0; v < r; v++) {
            var m = l[f * r + v]
            ;(p += m * u[v]), d.push(m)
          }
          if (p < 0 || p >= t.size / s)
            throw new Error(
              'Invalid indices: ' + d + ' does not index into ' + t.shape
            )
          for (var g = 0; g < s; g++) c.values[f * s + g] = h[p * s + g]
        }
        return c.toTensor().reshape(a)
      }),
      (o.prototype.scatterND = function (t, e, n) {
        var r = Zr(0, t, n),
          o = r.sliceRank,
          a = r.numUpdates,
          i = r.sliceSize,
          s = r.strides,
          u = r.outputSize,
          c = On(0)
        return this.scatter(t, e, n, u, i, a, o, s, c, !0)
      }),
      (o.prototype.fill = function (t, e, n) {
        var r = P((n = n || j(e)), k(t))
        return r.fill(e), Lt.makeTensor(r, t, n, this)
      }),
      (o.prototype.onesLike = function (t) {
        if ('string' === t.dtype)
          throw new Error('onesLike is not supported for string tensors')
        return this.fill(t.shape, 1, t.dtype)
      }),
      (o.prototype.zerosLike = function (t) {
        var e = P(t.dtype, k(t.shape))
        return this.makeOutput(e, t.shape, t.dtype)
      }),
      (o.prototype.linspace = function (t, e, n) {
        return Wo(t, e, n)
      }),
      (o.prototype.scatter = function (t, e, n, r, o, a, i, s, u, c) {
        var l = [r / o, o],
          h = this.readSync(t.dataId),
          f = this.readSync(e.dataId)
        if (0 === r) return Fn([], n, e.dtype)
        var d = new gt(l, e.dtype)
        d.values.fill(this.readSync(u.dataId)[0])
        for (var p = 0; p < a; p++) {
          for (var v = [], m = 0, g = 0; g < i; g++) {
            var y = h[p * i + g]
            v.push(y), (m += y * s[g])
          }
          if (m < 0 || m >= r / o)
            throw new Error(
              'Invalid indices: ' + v + ' does not index into ' + n
            )
          for (var x = 0; x < o; x++)
            c
              ? (d.values[m * o + x] += f[p * o + x])
              : (d.values[m * o + x] = 0 === e.rank ? f[0] : f[p * o + x])
        }
        return d.toTensor().reshape(n)
      }),
      o
    )
  })(bo)
  Lt.registerBackend(
    'cpu',
    function () {
      return new Kh()
    },
    1
  )
  for (
    var jh = 0,
      Xh = [
        {
          kernelName: 'NonMaxSuppressionV5',
          backendName: 'cpu',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = t.attrs,
              o = e,
              a = o.boxes,
              i = o.scores,
              s = r,
              u = s.maxOutputSize,
              c = s.iouThreshold,
              l = s.scoreThreshold,
              h = s.softNmsSigma,
              f = n
            Hh(a, 'NonMaxSuppressionWithScore')
            var d = Xo(
              f.data.get(a.dataId).values,
              f.data.get(i.dataId).values,
              u,
              c,
              l,
              h
            )
            return [d.selectedIndices, d.selectedScores]
          },
        },
        {
          kernelName: 'Square',
          backendName: 'cpu',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = e.x,
              o = n
            Hh(r, 'square')
            for (
              var a = o.data.get(r.dataId).values,
                i = new Float32Array(a.length),
                s = 0;
              s < a.length;
              ++s
            ) {
              var u = a[s]
              i[s] = u * u
            }
            return {
              dataId: o.write(i, r.shape, r.dtype),
              shape: r.shape,
              dtype: r.dtype,
            }
          },
        },
        {
          kernelName: eu,
          backendName: 'cpu',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = e,
              o = r.a,
              a = r.b,
              i = n
            Hh([o, a], eu)
            var s = i.data.get(o.dataId).values,
              u = i.data.get(a.dataId).values,
              c = (function (t, e, n, r, o, a) {
                var i = Ro(t, e),
                  s = i.length,
                  u = $(i),
                  c = B(o, k(i)),
                  l = t.length,
                  h = e.length,
                  f = $(t),
                  d = $(e),
                  p = Co(t, i),
                  v = Co(e, i)
                if (p.length + v.length === 0)
                  for (var m = 0; m < c.length; ++m)
                    c[m] = a(n[m % n.length], r[m % r.length])
                else {
                  var g = function (t) {
                    var e = it(t, s, u),
                      o = e.slice(-l)
                    p.forEach(function (t) {
                      return (o[t] = 0)
                    })
                    var i = at(o, l, f),
                      m = e.slice(-h)
                    v.forEach(function (t) {
                      return (m[t] = 0)
                    })
                    var g = at(m, h, d)
                    c[t] = a(n[i], r[g])
                  }
                  for (m = 0; m < c.length; ++m) g(m)
                }
                return [c, i]
              })(o.shape, a.shape, s, u, o.dtype, function (t, e) {
                var n = t - e
                return n * n
              }),
              l = c[0],
              h = c[1]
            return { dataId: i.write(l, h, o.dtype), shape: h, dtype: o.dtype }
          },
        },
      ];
    jh < Xh.length;
    jh++
  ) {
    d(Xh[jh])
  }
  var Yh,
    $h = function (t) {
      this.variableNames = ['A']
      var e = ua(),
        n = t[0],
        r = t[1]
      ;(this.outputShape = t),
        (this.userCode =
          '\n      void main() {\n        ivec3 coords = getOutputCoords();\n        int texR = coords[0];\n        int texC = coords[1];\n        int depth = coords[2];\n        vec2 uv = (vec2(texC, texR) + halfCR) / vec2(' +
          r +
          '.0, ' +
          n +
          '.0);\n\n        vec4 values = ' +
          e.texture2D +
          '(A, uv);\n        float value;\n        if (depth == 0) {\n          value = values.r;\n        } else if (depth == 1) {\n          value = values.g;\n        } else if (depth == 2) {\n          value = values.b;\n        } else if (depth == 3) {\n          value = values.a;\n        }\n\n        setOutput(floor(value * 255.0 + 0.5));\n      }\n    ')
    },
    Qh = function (t) {
      ;(this.variableNames = ['A']),
        (this.packedInputs = !1),
        (this.packedOutput = !0)
      var e = ua(),
        n = t[0],
        r = t[1]
      ;(this.outputShape = t),
        (this.userCode =
          '\n      void main() {\n        ivec3 coords = getOutputCoords();\n        int texR = coords[0];\n        int texC = coords[1];\n        int depth = coords[2];\n\n        vec4 result = vec4(0.);\n\n        for(int row=0; row<=1; row++) {\n          for(int col=0; col<=1; col++) {\n            texC = coords[1] + row;\n            depth = coords[2] + col;\n\n            vec2 uv = (vec2(texC, texR) + halfCR) /\n                       vec2(' +
          r +
          '.0, ' +
          n +
          '.0);\n            vec4 values = ' +
          e.texture2D +
          '(A, uv);\n            float value;\n            if (depth == 0) {\n              value = values.r;\n            } else if (depth == 1) {\n              value = values.g;\n            } else if (depth == 2) {\n              value = values.b;\n            } else if (depth == 3) {\n              value = values.a;\n            }\n\n            result[row * 2 + col] = floor(value * 255.0 + 0.5);\n          }\n        }\n\n        ' +
          e.output +
          ' = result;\n      }\n    ')
    }
  for (
    var Jh = 0,
      Zh = [
        {
          kernelName: 'FromPixels',
          backendName: 'webgl',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = t.attrs,
              o = e.pixels,
              a = r.numChannels,
              s =
                'undefined' != typeof HTMLVideoElement &&
                o instanceof HTMLVideoElement,
              u =
                'undefined' != typeof HTMLImageElement &&
                o instanceof HTMLImageElement,
              c = s ? [o.videoWidth, o.videoHeight] : [o.width, o.height],
              l = c[0],
              h = c[1],
              f = [h, l],
              d = [h, l, a]
            ;(u || s) &&
              (null == Yh &&
                (Yh = document.createElement('canvas').getContext('2d')),
              (Yh.canvas.width = l),
              (Yh.canvas.height = h),
              Yh.drawImage(o, 0, 0, l, h),
              (o = Yh.canvas))
            var p = n.makeTensorInfo(f, 'int32')
            ;(n.texData.get(p.dataId).usage = zt.PIXELS),
              n.gpgpu.uploadPixelDataToTexture(n.getTexture(p.dataId), o)
            var v = i().getBool('WEBGL_PACK') ? new Qh(d) : new $h(d),
              m = n.runWebGLProgram(v, [p], 'int32')
            return n.disposeData(p.dataId), m
          },
        },
        {
          kernelName: 'NonMaxSuppressionV5',
          backendName: 'webgl',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = t.attrs
            dn(
              'tf.nonMaxSuppression() in webgl locks the UI thread. Call tf.nonMaxSuppressionAsync() instead'
            )
            var o = e,
              a = o.boxes,
              i = o.scores,
              s = r,
              u = s.maxOutputSize,
              c = s.iouThreshold,
              l = s.scoreThreshold,
              h = s.softNmsSigma,
              f = n,
              d = Xo(f.readSync(a.dataId), f.readSync(i.dataId), u, c, l, h)
            return [d.selectedIndices, d.selectedScores]
          },
        },
        {
          kernelName: 'Square',
          backendName: 'webgl',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = e.x,
              o = n,
              a = new Cs(r.shape, 'return x * x;')
            return o.runWebGLProgram(a, [r], r.dtype)
          },
        },
        {
          kernelName: eu,
          backendName: 'webgl',
          kernelFunc: function (t) {
            var e = t.inputs,
              n = t.backend,
              r = e,
              o = r.a,
              a = r.b,
              s = n,
              u = i().getBool('WEBGL_PACK_BINARY_OPERATIONS')
                ? new La('return (a - b) * (a - b);', o.shape, a.shape)
                : new Ba('return (a - b) * (a - b);', o.shape, a.shape)
            return s.compileAndRun(u, [o, a])
          },
        },
      ];
    Jh < Zh.length;
    Jh++
  ) {
    d(Zh[Jh])
  }
  for (
    var tf = 0,
      ef = [
        {
          kernelName: 'Square',
          gradFunc: function (t, e) {
            var n = e[0]
            return {
              x: function () {
                return t.mul(n.toFloat().mul(2))
              },
            }
          },
        },
        {
          kernelName: eu,
          gradFunc: function (t, e) {
            var n = e[0],
              r = e[1],
              o = On(2)
            return {
              a: function () {
                return gc(t, gc(o, Cc(n, r)))
              },
              b: function () {
                return gc(t, gc(o, Cc(r, n)))
              },
            }
          },
        },
      ];
    tf < ef.length;
    tf++
  ) {
    p(ef[tf])
  }
  var nf = (function () {
    function t() {}
    return (
      (t.prototype.fetch = function (t, e) {
        return fetch(t, e)
      }),
      (t.prototype.now = function () {
        return performance.now()
      }),
      (t.prototype.encode = function (t, e) {
        if ('utf-8' !== e && 'utf8' !== e)
          throw new Error("Browser's encoder only supports utf-8, but got " + e)
        return (
          null == this.textEncoder && (this.textEncoder = new TextEncoder()),
          this.textEncoder.encode(t)
        )
      }),
      (t.prototype.decode = function (t, e) {
        return new TextDecoder(e).decode(t)
      }),
      t
    )
  })()
  i().get('IS_BROWSER') && i().setPlatform('browser', new nf())
  var rf,
    of = function () {
      return require('node-fetch')
    },
    af = (function () {
      function t() {
        ;(this.util = require('util')),
          (this.textEncoder = new this.util.TextEncoder())
      }
      return (
        (t.prototype.fetch = function (t, e) {
          return null != i().global.fetch
            ? i().global.fetch(t, e)
            : (null == rf && (rf = of()), rf(t, e))
        }),
        (t.prototype.now = function () {
          var t = process.hrtime()
          return 1e3 * t[0] + t[1] / 1e6
        }),
        (t.prototype.encode = function (t, e) {
          if ('utf-8' !== e && 'utf8' !== e)
            throw new Error(
              'Node built-in encoder only supports utf-8, but got ' + e
            )
          return this.textEncoder.encode(t)
        }),
        (t.prototype.decode = function (t, e) {
          return 0 === t.length ? '' : new this.util.TextDecoder(e).decode(t)
        }),
        t
      )
    })()
  i().get('IS_NODE') && i().setPlatform('node', new af())
  var sf = { float32: 4, int32: 4, uint16: 2, uint8: 1, bool: 1 },
    uf = 4
  function cf(t, e) {
    for (
      var n = {},
        r = 0,
        o = function (e) {
          var o = e.name,
            a = e.dtype,
            i = e.shape,
            s = k(i),
            u = void 0
          if (('quantization' in e)) {
            var c = e.quantization
            if ('uint8' !== c.dtype && 'uint16' !== c.dtype)
              throw new Error(
                'Weight ' +
                  e.name +
                  ' has unknown quantization dtype ' +
                  c.dtype +
                  ". Supported quantization dtypes are: 'uint8' and 'uint16'."
              )
            var l = sf[c.dtype],
              h = t.slice(r, r + s * l),
              f = 'uint8' === c.dtype ? new Uint8Array(h) : new Uint16Array(h)
            if ('float32' === a)
              u = Float32Array.from(f, function (t) {
                return t * c.scale + c.min
              })
            else {
              if ('int32' !== a)
                throw new Error("Unsupported dtype in weight '" + o + "': " + a)
              u = Int32Array.from(f, function (t) {
                return Math.round(t * c.scale + c.min)
              })
            }
            r += s * l
          } else if ('string' === a) {
            var d = k(e.shape)
            u = []
            for (var p = 0; p < d; p++) {
              var v = new Uint32Array(t.slice(r, r + uf))[0]
              r += uf
              var m = new Uint8Array(t.slice(r, r + v))
              u.push(m), (r += v)
            }
          } else {
            var g = sf[a]
            h = t.slice(r, r + s * g)
            if ('float32' === a) u = new Float32Array(h)
            else if ('int32' === a) u = new Int32Array(h)
            else {
              if ('bool' !== a)
                throw new Error("Unsupported dtype in weight '" + o + "': " + a)
              u = new Uint8Array(h)
            }
            r += s * g
          }
          n[o] = Fn(u, i, a)
        },
        a = 0,
        i = e;
      a < i.length;
      a++
    ) {
      o(i[a])
    }
    return n
  }
  function lf(t) {
    if (null === t) throw new Error('Invalid input value: ' + JSON.stringify(t))
    var e = 0,
      n = []
    t.forEach(function (t) {
      if (
        ((e += t.byteLength),
        n.push(t.byteLength === t.buffer.byteLength ? t : new t.constructor(t)),
        !(
          t instanceof Float32Array ||
          t instanceof Int32Array ||
          t instanceof Uint8Array
        ))
      )
        throw new Error('Unsupported TypedArray subtype: ' + t.constructor.name)
    })
    var r = new Uint8Array(e),
      o = 0
    return (
      n.forEach(function (t) {
        r.set(new Uint8Array(t.buffer), o), (o += t.byteLength)
      }),
      r.buffer
    )
  }
  var hf =
    'undefined' != typeof Buffer &&
    ('undefined' == typeof Blob ||
      'undefined' == typeof atob ||
      'undefined' == typeof btoa)
  function ff(t) {
    return hf ? Buffer.byteLength(t) : new Blob([t]).size
  }
  function df(t) {
    var e = 0
    t.forEach(function (t) {
      e += t.byteLength
    })
    var n = new Uint8Array(e),
      r = 0
    return (
      t.forEach(function (t) {
        n.set(new Uint8Array(t), r), (r += t.byteLength)
      }),
      n.buffer
    )
  }
  function pf(t) {
    for (t = t.trim(); t.endsWith('/'); ) t = t.slice(0, t.length - 1)
    var e = t.split('/')
    return e[e.length - 1]
  }
  function vf(t) {
    if (t.modelTopology instanceof ArrayBuffer)
      throw new Error('Expected JSON model topology, received ArrayBuffer.')
    return {
      dateSaved: new Date(),
      modelTopologyType: 'JSON',
      modelTopologyBytes:
        null == t.modelTopology ? 0 : ff(JSON.stringify(t.modelTopology)),
      weightSpecsBytes:
        null == t.weightSpecs ? 0 : ff(JSON.stringify(t.weightSpecs)),
      weightDataBytes: null == t.weightData ? 0 : t.weightData.byteLength,
    }
  }
  var mf = (function () {
      function t() {
        ;(this.saveRouters = []), (this.loadRouters = [])
      }
      return (
        (t.getInstance = function () {
          return null == t.instance && (t.instance = new t()), t.instance
        }),
        (t.registerSaveRouter = function (e) {
          t.getInstance().saveRouters.push(e)
        }),
        (t.registerLoadRouter = function (e) {
          t.getInstance().loadRouters.push(e)
        }),
        (t.getSaveHandlers = function (e) {
          return t.getHandlers(e, 'save')
        }),
        (t.getLoadHandlers = function (e, n) {
          return t.getHandlers(e, 'load', n)
        }),
        (t.getHandlers = function (e, n, r) {
          var o = []
          return (
            ('load' === n
              ? t.getInstance().loadRouters
              : t.getInstance().saveRouters
            ).forEach(function (t) {
              var n = t(e, r)
              null !== n && o.push(n)
            }),
            o
          )
        }),
        t
      )
    })(),
    gf = '://',
    yf = (function () {
      function t() {
        this.managers = {}
      }
      return (
        (t.getInstance = function () {
          return null == t.instance && (t.instance = new t()), t.instance
        }),
        (t.registerManager = function (e, n) {
          C(null != e, function () {
            return 'scheme must not be undefined or null.'
          }),
            e.endsWith(gf) && (e = e.slice(0, e.indexOf(gf))),
            C(e.length > 0, function () {
              return 'scheme must not be an empty string.'
            })
          var r = t.getInstance()
          C(null == r.managers[e], function () {
            return (
              "A model store manager is already registered for scheme '" +
              e +
              "'."
            )
          }),
            (r.managers[e] = n)
        }),
        (t.getManager = function (t) {
          var e = this.getInstance().managers[t]
          if (null == e)
            throw new Error("Cannot find model manager for scheme '" + t + "'")
          return e
        }),
        (t.getSchemes = function () {
          return Object.keys(this.getInstance().managers)
        }),
        t
      )
    })()
  function xf(t) {
    if (-1 === t.indexOf(gf))
      throw new Error(
        'The url string provided does not contain a scheme. Supported schemes are: ' +
          yf.getSchemes().join(',')
      )
    return { scheme: t.split(gf)[0], path: t.split(gf)[1] }
  }
  function bf(t, e, o) {
    return (
      void 0 === o && (o = !1),
      n(this, void 0, void 0, function () {
        var n, a, i, s, u, c, l, h, f
        return r(this, function (r) {
          switch (r.label) {
            case 0:
              return (
                C(t !== e, function () {
                  return "Old path and new path are the same: '" + t + "'"
                }),
                C((n = mf.getLoadHandlers(t)).length > 0, function () {
                  return (
                    'Copying failed because no load handler is found for source URL ' +
                    t +
                    '.'
                  )
                }),
                C(n.length < 2, function () {
                  return (
                    'Copying failed because more than one (' +
                    n.length +
                    ') load handlers for source URL ' +
                    t +
                    '.'
                  )
                }),
                (a = n[0]),
                C((i = mf.getSaveHandlers(e)).length > 0, function () {
                  return (
                    'Copying failed because no save handler is found for destination URL ' +
                    e +
                    '.'
                  )
                }),
                C(i.length < 2, function () {
                  return (
                    'Copying failed because more than one (' +
                    n.length +
                    ') save handlers for destination URL ' +
                    e +
                    '.'
                  )
                }),
                (s = i[0]),
                (u = xf(t).scheme),
                (c = xf(t).path),
                (l = u === xf(t).scheme),
                [4, a.load()]
              )
            case 1:
              return (
                (h = r.sent()),
                o && l ? [4, yf.getManager(u).removeModel(c)] : [3, 3]
              )
            case 2:
              r.sent(), (r.label = 3)
            case 3:
              return [4, s.save(h)]
            case 4:
              return (
                (f = r.sent()),
                !o || l ? [3, 6] : [4, yf.getManager(u).removeModel(c)]
              )
            case 5:
              r.sent(), (r.label = 6)
            case 6:
              return [2, f.modelArtifactsInfo]
          }
        })
      })
    )
  }
  var wf = 'models_store',
    Cf = 'model_info_store'
  function Ef() {
    if (!i().getBool('IS_BROWSER'))
      throw new Error(
        'Failed to obtain IndexedDB factory because the current environmentis not a web browser.'
      )
    var t = window || self,
      e =
        t.indexedDB ||
        t.mozIndexedDB ||
        t.webkitIndexedDB ||
        t.msIndexedDB ||
        t.shimIndexedDB
    if (null == e)
      throw new Error(
        'The current browser does not appear to support IndexedDB.'
      )
    return e
  }
  function Rf(t) {
    var e = t.result
    e.createObjectStore(wf, { keyPath: 'modelPath' }),
      e.createObjectStore(Cf, { keyPath: 'modelPath' })
  }
  var If = (function () {
      function t(t) {
        if (((this.indexedDB = Ef()), null == t || !t))
          throw new Error(
            'For IndexedDB, modelPath must not be null, undefined or empty.'
          )
        this.modelPath = t
      }
      return (
        (t.prototype.save = function (t) {
          return n(this, void 0, void 0, function () {
            return r(this, function (e) {
              if (t.modelTopology instanceof ArrayBuffer)
                throw new Error(
                  'BrowserLocalStorage.save() does not support saving model topology in binary formats yet.'
                )
              return [2, this.databaseAction(this.modelPath, t)]
            })
          })
        }),
        (t.prototype.load = function () {
          return n(this, void 0, void 0, function () {
            return r(this, function (t) {
              return [2, this.databaseAction(this.modelPath)]
            })
          })
        }),
        (t.prototype.databaseAction = function (t, e) {
          var n = this
          return new Promise(function (t, r) {
            var o = n.indexedDB.open('tensorflowjs', 1)
            ;(o.onupgradeneeded = function () {
              return Rf(o)
            }),
              (o.onsuccess = function () {
                var a = o.result
                if (null == e) {
                  var i = a.transaction(wf, 'readonly'),
                    s = i.objectStore(wf).get(n.modelPath)
                  ;(s.onsuccess = function () {
                    if (null == s.result)
                      return (
                        a.close(),
                        r(
                          new Error(
                            "Cannot find model with path '" +
                              n.modelPath +
                              "' in IndexedDB."
                          )
                        )
                      )
                    t(s.result.modelArtifacts)
                  }),
                    (s.onerror = function (t) {
                      return a.close(), r(s.error)
                    }),
                    (i.oncomplete = function () {
                      return a.close()
                    })
                } else {
                  var u,
                    c = vf(e),
                    l = a.transaction(Cf, 'readwrite'),
                    h = l.objectStore(Cf),
                    f = h.put({ modelPath: n.modelPath, modelArtifactsInfo: c })
                  ;(f.onsuccess = function () {
                    var o = (u = a.transaction(wf, 'readwrite'))
                      .objectStore(wf)
                      .put({
                        modelPath: n.modelPath,
                        modelArtifacts: e,
                        modelArtifactsInfo: c,
                      })
                    ;(o.onsuccess = function () {
                      return t({ modelArtifactsInfo: c })
                    }),
                      (o.onerror = function (t) {
                        var e = (h = l.objectStore(Cf)).delete(n.modelPath)
                        ;(e.onsuccess = function () {
                          return a.close(), r(o.error)
                        }),
                          (e.onerror = function (t) {
                            return a.close(), r(o.error)
                          })
                      })
                  }),
                    (f.onerror = function (t) {
                      return a.close(), r(f.error)
                    }),
                    (l.oncomplete = function () {
                      null == u
                        ? a.close()
                        : (u.oncomplete = function () {
                            return a.close()
                          })
                    })
                }
              }),
              (o.onerror = function (t) {
                return r(o.error)
              })
          })
        }),
        (t.URL_SCHEME = 'indexeddb://'),
        t
      )
    })(),
    kf = function (t) {
      return i().getBool('IS_BROWSER') &&
        !Array.isArray(t) &&
        t.startsWith(If.URL_SCHEME)
        ? ((e = t.slice(If.URL_SCHEME.length)), new If(e))
        : null
      var e
    }
  mf.registerSaveRouter(kf), mf.registerLoadRouter(kf)
  var Sf = (function () {
    function t() {
      this.indexedDB = Ef()
    }
    return (
      (t.prototype.listModels = function () {
        return n(this, void 0, void 0, function () {
          var t = this
          return r(this, function (e) {
            return [
              2,
              new Promise(function (e, n) {
                var r = t.indexedDB.open('tensorflowjs', 1)
                ;(r.onupgradeneeded = function () {
                  return Rf(r)
                }),
                  (r.onsuccess = function () {
                    var t = r.result,
                      o = t.transaction(Cf, 'readonly'),
                      a = o.objectStore(Cf).getAll()
                    ;(a.onsuccess = function () {
                      for (var t = {}, n = 0, r = a.result; n < r.length; n++) {
                        var o = r[n]
                        t[o.modelPath] = o.modelArtifactsInfo
                      }
                      e(t)
                    }),
                      (a.onerror = function (e) {
                        return t.close(), n(a.error)
                      }),
                      (o.oncomplete = function () {
                        return t.close()
                      })
                  }),
                  (r.onerror = function (t) {
                    return n(r.error)
                  })
              }),
            ]
          })
        })
      }),
      (t.prototype.removeModel = function (t) {
        return n(this, void 0, void 0, function () {
          var e = this
          return r(this, function (n) {
            var r
            return (
              (t = (r = t).startsWith(If.URL_SCHEME)
                ? r.slice(If.URL_SCHEME.length)
                : r),
              [
                2,
                new Promise(function (n, r) {
                  var o = e.indexedDB.open('tensorflowjs', 1)
                  ;(o.onupgradeneeded = function () {
                    return Rf(o)
                  }),
                    (o.onsuccess = function () {
                      var e,
                        a = o.result,
                        i = a.transaction(Cf, 'readwrite'),
                        s = i.objectStore(Cf),
                        u = s.get(t)
                      ;(u.onsuccess = function () {
                        if (null == u.result)
                          return (
                            a.close(),
                            r(
                              new Error(
                                "Cannot find model with path '" +
                                  t +
                                  "' in IndexedDB."
                              )
                            )
                          )
                        var o = s.delete(t),
                          i = function () {
                            var o = (e = a.transaction(wf, 'readwrite'))
                              .objectStore(wf)
                              .delete(t)
                            ;(o.onsuccess = function () {
                              return n(u.result.modelArtifactsInfo)
                            }),
                              (o.onerror = function (t) {
                                return r(u.error)
                              })
                          }
                        ;(o.onsuccess = i),
                          (o.onerror = function (t) {
                            return i(), a.close(), r(u.error)
                          })
                      }),
                        (u.onerror = function (t) {
                          return a.close(), r(u.error)
                        }),
                        (i.oncomplete = function () {
                          null == e
                            ? a.close()
                            : (e.oncomplete = function () {
                                return a.close()
                              })
                        })
                    }),
                    (o.onerror = function (t) {
                      return r(o.error)
                    })
                }),
              ]
            )
          })
        })
      }),
      t
    )
  })()
  if (i().getBool('IS_BROWSER'))
    try {
      yf.registerManager(If.URL_SCHEME, new Sf())
    } catch (t) {}
  var Af = '/',
    Df = 'tensorflowjs_models',
    Tf = 'info',
    Nf = 'model_topology',
    Ff = 'weight_specs',
    _f = 'weight_data',
    Of = 'model_metadata'
  function Mf(t) {
    return {
      info: [Df, t, Tf].join(Af),
      topology: [Df, t, Nf].join(Af),
      weightSpecs: [Df, t, Ff].join(Af),
      weightData: [Df, t, _f].join(Af),
      modelMetadata: [Df, t, Of].join(Af),
    }
  }
  function Bf(t) {
    var e = t.split(Af)
    if (e.length < 3) throw new Error('Invalid key format: ' + t)
    return e.slice(1, e.length - 1).join(Af)
  }
  var Pf = (function () {
      function t(t) {
        if (
          !i().getBool('IS_BROWSER') ||
          'undefined' == typeof window ||
          void 0 === window.localStorage
        )
          throw new Error(
            'The current environment does not support local storage.'
          )
        if (((this.LS = window.localStorage), null == t || !t))
          throw new Error(
            'For local storage, modelPath must not be null, undefined or empty.'
          )
        ;(this.modelPath = t), (this.keys = Mf(this.modelPath))
      }
      return (
        (t.prototype.save = function (t) {
          return n(this, void 0, void 0, function () {
            var e, n, o
            return r(this, function (r) {
              if (t.modelTopology instanceof ArrayBuffer)
                throw new Error(
                  'BrowserLocalStorage.save() does not support saving model topology in binary formats yet.'
                )
              ;(e = JSON.stringify(t.modelTopology)),
                (n = JSON.stringify(t.weightSpecs)),
                (o = vf(t))
              try {
                return (
                  this.LS.setItem(this.keys.info, JSON.stringify(o)),
                  this.LS.setItem(this.keys.topology, e),
                  this.LS.setItem(this.keys.weightSpecs, n),
                  this.LS.setItem(
                    this.keys.weightData,
                    (function (t) {
                      if (hf) return Buffer.from(t).toString('base64')
                      for (
                        var e = new Uint8Array(t), n = '', r = 0, o = e.length;
                        r < o;
                        r++
                      )
                        n += String.fromCharCode(e[r])
                      return btoa(n)
                    })(t.weightData)
                  ),
                  this.LS.setItem(
                    this.keys.modelMetadata,
                    JSON.stringify({
                      format: t.format,
                      generatedBy: t.generatedBy,
                      convertedBy: t.convertedBy,
                      userDefinedMetadata: t.userDefinedMetadata,
                    })
                  ),
                  [2, { modelArtifactsInfo: o }]
                )
              } catch (t) {
                throw (
                  (this.LS.removeItem(this.keys.info),
                  this.LS.removeItem(this.keys.topology),
                  this.LS.removeItem(this.keys.weightSpecs),
                  this.LS.removeItem(this.keys.weightData),
                  this.LS.removeItem(this.keys.modelMetadata),
                  new Error(
                    "Failed to save model '" +
                      this.modelPath +
                      "' to local storage: size quota being exceeded is a possible cause of this failure: modelTopologyBytes=" +
                      o.modelTopologyBytes +
                      ', weightSpecsBytes=' +
                      o.weightSpecsBytes +
                      ', weightDataBytes=' +
                      o.weightDataBytes +
                      '.'
                  ))
                )
              }
              return [2]
            })
          })
        }),
        (t.prototype.load = function () {
          return n(this, void 0, void 0, function () {
            var t, e, n, o, a, i, s
            return r(this, function (r) {
              if (null == (t = JSON.parse(this.LS.getItem(this.keys.info))))
                throw new Error(
                  "In local storage, there is no model with name '" +
                    this.modelPath +
                    "'"
                )
              if ('JSON' !== t.modelTopologyType)
                throw new Error(
                  'BrowserLocalStorage does not support loading non-JSON model topology yet.'
                )
              if (
                ((e = {}),
                null == (n = JSON.parse(this.LS.getItem(this.keys.topology))))
              )
                throw new Error(
                  "In local storage, the topology of model '" +
                    this.modelPath +
                    "' is missing."
                )
              if (
                ((e.modelTopology = n),
                null ==
                  (o = JSON.parse(this.LS.getItem(this.keys.weightSpecs))))
              )
                throw new Error(
                  "In local storage, the weight specs of model '" +
                    this.modelPath +
                    "' are missing."
                )
              if (
                ((e.weightSpecs = o),
                null != (a = this.LS.getItem(this.keys.modelMetadata)) &&
                  ((i = JSON.parse(a)),
                  (e.format = i.format),
                  (e.generatedBy = i.generatedBy),
                  (e.convertedBy = i.convertedBy),
                  (e.userDefinedMetadata = i.userDefinedMetadata)),
                null == (s = this.LS.getItem(this.keys.weightData)))
              )
                throw new Error(
                  "In local storage, the binary weight values of model '" +
                    this.modelPath +
                    "' are missing."
                )
              return (
                (e.weightData = (function (t) {
                  if (hf) {
                    var e = Buffer.from(t, 'base64')
                    return e.buffer.slice(
                      e.byteOffset,
                      e.byteOffset + e.byteLength
                    )
                  }
                  for (
                    var n = atob(t), r = new Uint8Array(n.length), o = 0;
                    o < n.length;
                    ++o
                  )
                    r.set([n.charCodeAt(o)], o)
                  return r.buffer
                })(s)),
                [2, e]
              )
            })
          })
        }),
        (t.URL_SCHEME = 'localstorage://'),
        t
      )
    })(),
    Lf = function (t) {
      return i().getBool('IS_BROWSER') &&
        !Array.isArray(t) &&
        t.startsWith(Pf.URL_SCHEME)
        ? ((e = t.slice(Pf.URL_SCHEME.length)), new Pf(e))
        : null
      var e
    }
  mf.registerSaveRouter(Lf), mf.registerLoadRouter(Lf)
  var Wf = (function () {
    function t() {
      C(i().getBool('IS_BROWSER'), function () {
        return 'Current environment is not a web browser'
      }),
        C(
          'undefined' == typeof window || void 0 !== window.localStorage,
          function () {
            return 'Current browser does not appear to support localStorage'
          }
        ),
        (this.LS = window.localStorage)
    }
    return (
      (t.prototype.listModels = function () {
        return n(this, void 0, void 0, function () {
          var t, e, n, o, a, i
          return r(this, function (r) {
            for (
              t = {}, e = Df + Af, n = Af + Tf, o = 0;
              o < this.LS.length;
              ++o
            )
              (a = this.LS.key(o)).startsWith(e) &&
                a.endsWith(n) &&
                ((i = Bf(a)), (t[i] = JSON.parse(this.LS.getItem(a))))
            return [2, t]
          })
        })
      }),
      (t.prototype.removeModel = function (t) {
        return n(this, void 0, void 0, function () {
          var e, n
          return r(this, function (r) {
            var o
            if (
              ((t = (o = t).startsWith(Pf.URL_SCHEME)
                ? o.slice(Pf.URL_SCHEME.length)
                : o),
              (e = Mf(t)),
              null == this.LS.getItem(e.info))
            )
              throw new Error("Cannot find model at path '" + t + "'")
            return (
              (n = JSON.parse(this.LS.getItem(e.info))),
              this.LS.removeItem(e.info),
              this.LS.removeItem(e.topology),
              this.LS.removeItem(e.weightSpecs),
              this.LS.removeItem(e.weightData),
              [2, n]
            )
          })
        })
      }),
      t
    )
  })()
  if (i().getBool('IS_BROWSER'))
    try {
      yf.registerManager(Pf.URL_SCHEME, new Wf())
    } catch (t) {}
  var Uf = 'model',
    Vf = '.json',
    zf = '.weights.bin'
  function Gf(t) {
    return new Promise(function (t) {
      return setTimeout(t)
    }).then(t)
  }
  var Hf = (function () {
      function t(e) {
        if (!i().getBool('IS_BROWSER'))
          throw new Error(
            'browserDownloads() cannot proceed because the current environment is not a browser.'
          )
        e.startsWith(t.URL_SCHEME) && (e = e.slice(t.URL_SCHEME.length)),
          (null != e && 0 !== e.length) || (e = Uf),
          (this.modelTopologyFileName = e + Vf),
          (this.weightDataFileName = e + zf)
      }
      return (
        (t.prototype.save = function (t) {
          return n(this, void 0, void 0, function () {
            var e, n, o, a, i, s
            return r(this, function (r) {
              switch (r.label) {
                case 0:
                  if ('undefined' == typeof document)
                    throw new Error(
                      'Browser downloads are not supported in this environment since `document` is not present'
                    )
                  if (
                    ((e = window.URL.createObjectURL(
                      new Blob([t.weightData], {
                        type: 'application/octet-stream',
                      })
                    )),
                    !(t.modelTopology instanceof ArrayBuffer))
                  )
                    return [3, 1]
                  throw new Error(
                    'BrowserDownloads.save() does not support saving model topology in binary formats yet.'
                  )
                case 1:
                  return (
                    (n = [
                      {
                        paths: ['./' + this.weightDataFileName],
                        weights: t.weightSpecs,
                      },
                    ]),
                    (o = {
                      modelTopology: t.modelTopology,
                      format: t.format,
                      generatedBy: t.generatedBy,
                      convertedBy: t.convertedBy,
                      weightsManifest: n,
                    }),
                    (a = window.URL.createObjectURL(
                      new Blob([JSON.stringify(o)], {
                        type: 'application/json',
                      })
                    )),
                    ((i =
                      null == this.jsonAnchor
                        ? document.createElement('a')
                        : this.jsonAnchor).download =
                      this.modelTopologyFileName),
                    (i.href = a),
                    [
                      4,
                      Gf(function () {
                        return i.dispatchEvent(new MouseEvent('click'))
                      }),
                    ]
                  )
                case 2:
                  return (
                    r.sent(),
                    null == t.weightData
                      ? [3, 4]
                      : (((s =
                          null == this.weightDataAnchor
                            ? document.createElement('a')
                            : this.weightDataAnchor).download =
                          this.weightDataFileName),
                        (s.href = e),
                        [
                          4,
                          Gf(function () {
                            return s.dispatchEvent(new MouseEvent('click'))
                          }),
                        ])
                  )
                case 3:
                  r.sent(), (r.label = 4)
                case 4:
                  return [2, { modelArtifactsInfo: vf(t) }]
              }
            })
          })
        }),
        (t.URL_SCHEME = 'downloads://'),
        t
      )
    })(),
    qf = (function () {
      function t(t) {
        if (null == t || t.length < 1)
          throw new Error(
            'When calling browserFiles, at least 1 file is required, but received ' +
              t
          )
        this.files = t
      }
      return (
        (t.prototype.load = function () {
          return n(this, void 0, void 0, function () {
            var t,
              e,
              n = this
            return r(this, function (r) {
              return (
                (t = this.files[0]),
                (e = this.files.slice(1)),
                [
                  2,
                  new Promise(function (r, o) {
                    var a = new FileReader()
                    ;(a.onload = function (a) {
                      var i = JSON.parse(a.target.result),
                        s = i.modelTopology
                      if (null != s) {
                        0 === e.length && r({ modelTopology: s })
                        var u = i.weightsManifest
                        if (null != u) {
                          var c
                          try {
                            c = n.checkManifestAndWeightFiles(u, e)
                          } catch (t) {
                            return void o(t)
                          }
                          var l = [],
                            h = [],
                            f = []
                          u.forEach(function (t) {
                            t.paths.forEach(function (t) {
                              h.push(t), f.push(null)
                            }),
                              l.push.apply(l, t.weights)
                          }),
                            u.forEach(function (t) {
                              t.paths.forEach(function (t) {
                                var e = new FileReader()
                                ;(e.onload = function (e) {
                                  var n = e.target.result,
                                    o = h.indexOf(t)
                                  ;(f[o] = n),
                                    -1 === f.indexOf(null) &&
                                      r({
                                        modelTopology: s,
                                        weightSpecs: l,
                                        weightData: df(f),
                                        format: i.format,
                                        generatedBy: i.generatedBy,
                                        convertedBy: i.convertedBy,
                                        userDefinedMetadata:
                                          i.userDefinedMetadata,
                                      })
                                }),
                                  (e.onerror = function (e) {
                                    return o(
                                      "Failed to weights data from file of path '" +
                                        t +
                                        "'."
                                    )
                                  }),
                                  e.readAsArrayBuffer(c[t])
                              })
                            })
                        } else
                          o(
                            new Error(
                              'weightManifest field is missing from file ' +
                                t.name
                            )
                          )
                      } else
                        o(
                          new Error(
                            'modelTopology field is missing from file ' + t.name
                          )
                        )
                    }),
                      (a.onerror = function (e) {
                        return o(
                          "Failed to read model topology and weights manifest JSON from file '" +
                            t.name +
                            "'. BrowserFiles supports loading Keras-style tf.Model artifacts only."
                        )
                      }),
                      a.readAsText(t)
                  }),
                ]
              )
            })
          })
        }),
        (t.prototype.checkManifestAndWeightFiles = function (t, e) {
          for (
            var n = [],
              r = e.map(function (t) {
                return pf(t.name)
              }),
              o = {},
              a = 0,
              i = t;
            a < i.length;
            a++
          ) {
            i[a].paths.forEach(function (t) {
              var a = pf(t)
              if (-1 !== n.indexOf(a))
                throw new Error(
                  "Duplicate file basename found in weights manifest: '" +
                    a +
                    "'"
                )
              if ((n.push(a), -1 === r.indexOf(a)))
                throw new Error(
                  "Weight file with basename '" + a + "' is not provided."
                )
              o[t] = e[r.indexOf(a)]
            })
          }
          if (n.length !== e.length)
            throw new Error(
              'Mismatch in the number of files in weights manifest (' +
                n.length +
                ') and the number of weight files provided (' +
                e.length +
                ').'
            )
          return o
        }),
        t
      )
    })()
  function Kf(t, e, n, r) {
    !(function (t) {
      C(null != t && Array.isArray(t) && t.length > 0, function () {
        return 'promises must be a none empty array'
      })
    })(t),
      (function (t, e) {
        C(t >= 0 && t <= 1, function () {
          return (
            'Progress fraction must be in range [0, 1], but got startFraction ' +
            t
          )
        }),
          C(e >= 0 && e <= 1, function () {
            return (
              'Progress fraction must be in range [0, 1], but got endFraction ' +
              e
            )
          }),
          C(e >= t, function () {
            return (
              'startFraction must be no more than endFraction, but got startFraction ' +
              t +
              ' and endFraction ' +
              e
            )
          })
      })((n = null == n ? 0 : n), (r = null == r ? 1 : r))
    var o = 0
    return Promise.all(
      t.map(function (a) {
        return (
          a.then(function (a) {
            var i = n + (++o / t.length) * (r - n)
            return e(i), a
          }),
          a
        )
      })
    )
  }
  function jf(t, e) {
    return n(this, void 0, void 0, function () {
      var n, o, a, s, u, c, l, h, f
      return r(this, function (r) {
        switch (r.label) {
          case 0:
            return (
              null == e && (e = {}),
              (n = null == e.fetchFunc ? i().platform.fetch : e.fetchFunc),
              (o = t.map(function (t) {
                return n(t, e.requestInit, { isBinary: !0 })
              })),
              (a = 0),
              (s = 0.5),
              null != e.onProgress ? [3, 2] : [4, Promise.all(o)]
            )
          case 1:
            return (u = r.sent()), [3, 4]
          case 2:
            return [4, Kf(o, e.onProgress, a, s)]
          case 3:
            ;(u = r.sent()), (r.label = 4)
          case 4:
            return (
              (c = u.map(function (t) {
                return t.arrayBuffer()
              })),
              (l = 0.5),
              (h = 1),
              null != e.onProgress ? [3, 6] : [4, Promise.all(c)]
            )
          case 5:
            return (f = r.sent()), [3, 8]
          case 6:
            return [4, Kf(c, e.onProgress, l, h)]
          case 7:
            ;(f = r.sent()), (r.label = 8)
          case 8:
            return [2, f]
        }
      })
    })
  }
  function Xf(t) {
    var e = this
    return function (o, a, i) {
      return (
        void 0 === a && (a = ''),
        n(e, void 0, void 0, function () {
          var e, n, s, u, c, l, h, f, d, p
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                if (
                  ((e = o.map(function () {
                    return !1
                  })),
                  (n = {}),
                  (s =
                    null != i
                      ? i.map(function () {
                          return !1
                        })
                      : []),
                  (u = []),
                  o.forEach(function (t, r) {
                    var o = 0
                    t.weights.forEach(function (t) {
                      var a =
                          'quantization' in t ? t.quantization.dtype : t.dtype,
                        c = sf[a] * k(t.shape),
                        l = function () {
                          ;(e[r] = !0),
                            null == n[r] && (n[r] = []),
                            n[r].push({
                              manifestEntry: t,
                              groupOffset: o,
                              sizeBytes: c,
                            })
                        }
                      null != i
                        ? i.forEach(function (e, n) {
                            e === t.name && (l(), (s[n] = !0))
                          })
                        : l(),
                        u.push(t.name),
                        (o += c)
                    })
                  }),
                  !s.every(function (t) {
                    return t
                  }))
                )
                  throw (
                    ((c = i.filter(function (t, e) {
                      return !s[e]
                    })),
                    new Error(
                      'Could not find weights in manifest with names: ' +
                        c.join(', ') +
                        '. \nManifest JSON has weights with names: ' +
                        u.join(', ') +
                        '.'
                    ))
                  )
                return (
                  (l = e.reduce(function (t, e, n) {
                    return e && t.push(n), t
                  }, [])),
                  (h = []),
                  l.forEach(function (t) {
                    o[t].paths.forEach(function (t) {
                      var e = a + (a.endsWith('/') ? '' : '/') + t
                      h.push(e)
                    })
                  }),
                  [4, t(h)]
                )
              case 1:
                return (
                  (f = r.sent()),
                  (d = {}),
                  (p = 0),
                  l.forEach(function (t) {
                    for (var e = o[t].paths.length, r = 0, a = 0; a < e; a++)
                      r += f[p + a].byteLength
                    for (
                      var i = new ArrayBuffer(r),
                        s = new Uint8Array(i),
                        u = 0,
                        c = 0;
                      c < e;
                      c++
                    ) {
                      var l = new Uint8Array(f[p + c])
                      s.set(l, u), (u += l.byteLength)
                    }
                    n[t].forEach(function (t) {
                      var e = cf(
                        i.slice(t.groupOffset, t.groupOffset + t.sizeBytes),
                        [t.manifestEntry]
                      )
                      for (var n in e) d[n] = e[n]
                    }),
                      (p += e)
                  }),
                  [2, d]
                )
            }
          })
        })
      )
    }
  }
  mf.registerSaveRouter(function (t) {
    return i().getBool('IS_BROWSER') &&
      !Array.isArray(t) &&
      t.startsWith(Hf.URL_SCHEME)
      ? (function (t) {
          void 0 === t && (t = 'model')
          return new Hf(t)
        })(t.slice(Hf.URL_SCHEME.length))
      : null
  })
  var Yf = (function () {
    function t(t, e) {
      if (
        ((this.DEFAULT_METHOD = 'POST'),
        null == e && (e = {}),
        (this.weightPathPrefix = e.weightPathPrefix),
        (this.onProgress = e.onProgress),
        null != e.fetchFunc
          ? (C('function' == typeof e.fetchFunc, function () {
              return 'Must pass a function that matches the signature of `fetch` (see https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API)'
            }),
            (this.fetch = e.fetchFunc))
          : (this.fetch = i().platform.fetch),
        C(null != t && t.length > 0, function () {
          return 'URL path for http must not be null, undefined or empty.'
        }),
        Array.isArray(t) &&
          C(2 === t.length, function () {
            return (
              'URL paths for http must have a length of 2, (actual length is ' +
              t.length +
              ').'
            )
          }),
        (this.path = t),
        null != e.requestInit && null != e.requestInit.body)
      )
        throw new Error(
          'requestInit is expected to have no pre-existing body, but has one.'
        )
      this.requestInit = e.requestInit || {}
    }
    return (
      (t.prototype.save = function (t) {
        return n(this, void 0, void 0, function () {
          var e, n, o, a
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                if (t.modelTopology instanceof ArrayBuffer)
                  throw new Error(
                    'BrowserHTTPRequest.save() does not support saving model topology in binary formats yet.'
                  )
                return (
                  ((e = Object.assign(
                    { method: this.DEFAULT_METHOD },
                    this.requestInit
                  )).body = new FormData()),
                  (n = [
                    { paths: ['./model.weights.bin'], weights: t.weightSpecs },
                  ]),
                  (o = {
                    modelTopology: t.modelTopology,
                    format: t.format,
                    generatedBy: t.generatedBy,
                    convertedBy: t.convertedBy,
                    userDefinedMetadata: t.userDefinedMetadata,
                    weightsManifest: n,
                  }),
                  e.body.append(
                    'model.json',
                    new Blob([JSON.stringify(o)], { type: 'application/json' }),
                    'model.json'
                  ),
                  null != t.weightData &&
                    e.body.append(
                      'model.weights.bin',
                      new Blob([t.weightData], {
                        type: 'application/octet-stream',
                      }),
                      'model.weights.bin'
                    ),
                  [4, this.fetch(this.path, e)]
                )
              case 1:
                if ((a = r.sent()).ok)
                  return [2, { modelArtifactsInfo: vf(t), responses: [a] }]
                throw new Error(
                  'BrowserHTTPRequest.save() failed due to HTTP response status ' +
                    a.status +
                    '.'
                )
            }
          })
        })
      }),
      (t.prototype.load = function () {
        return n(this, void 0, void 0, function () {
          var t, e, n, o, a, i, s, u, c, l, h, f
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                return [4, this.fetch(this.path, this.requestInit)]
              case 1:
                if (!(t = r.sent()).ok)
                  throw new Error(
                    'Request to ' +
                      this.path +
                      ' failed with status code ' +
                      t.status +
                      '. Please verify this URL points to the model JSON of the model to load.'
                  )
                r.label = 2
              case 2:
                return r.trys.push([2, 4, , 5]), [4, t.json()]
              case 3:
                return (e = r.sent()), [3, 5]
              case 4:
                throw (
                  (r.sent(),
                  (n =
                    'Failed to parse model JSON of response from ' +
                    this.path +
                    '.'),
                  this.path.endsWith('.pb')
                    ? (n +=
                        " Your path contains a .pb file extension. Support for .pb models have been removed in TensorFlow.js 1.0 in favor of .json models. You can re-convert your Python TensorFlow model using the TensorFlow.js 1.0 conversion scripts or you can convert your.pb models with the 'pb2json'NPM script in the tensorflow/tfjs-converter repository.")
                    : (n +=
                        ' Please make sure the server is serving valid JSON for this request.'),
                  new Error(n))
                )
              case 5:
                if (
                  ((o = e.modelTopology),
                  (a = e.weightsManifest),
                  (i = e.generatedBy),
                  (s = e.convertedBy),
                  (u = e.format),
                  (c = e.userDefinedMetadata),
                  null == o && null == a)
                )
                  throw new Error(
                    'The JSON from HTTP path ' +
                      this.path +
                      ' contains neither model topology or manifest for weights.'
                  )
                return null == a ? [3, 7] : [4, this.loadWeights(a)]
              case 6:
                ;(f = r.sent()), (l = f[0]), (h = f[1]), (r.label = 7)
              case 7:
                return [
                  2,
                  {
                    modelTopology: o,
                    weightSpecs: l,
                    weightData: h,
                    userDefinedMetadata: c,
                    generatedBy: i,
                    convertedBy: s,
                    format: u,
                  },
                ]
            }
          })
        })
      }),
      (t.prototype.loadWeights = function (t) {
        return n(this, void 0, void 0, function () {
          var e, n, o, a, i, s, u, c, l, h, f
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                for (
                  e = Array.isArray(this.path) ? this.path[1] : this.path,
                    n = (function (t) {
                      var e = t.lastIndexOf('/'),
                        n = t.lastIndexOf('?'),
                        r = t.substring(0, e),
                        o = n > e ? t.substring(n) : ''
                      return [r + '/', o]
                    })(e),
                    o = n[0],
                    a = n[1],
                    i = this.weightPathPrefix || o,
                    s = [],
                    u = 0,
                    c = t;
                  u < c.length;
                  u++
                )
                  (l = c[u]), s.push.apply(s, l.weights)
                return (
                  (h = []),
                  t.forEach(function (t) {
                    t.paths.forEach(function (t) {
                      h.push(i + t + a)
                    })
                  }),
                  [
                    4,
                    jf(h, {
                      requestInit: this.requestInit,
                      fetchFunc: this.fetch,
                      onProgress: this.onProgress,
                    }),
                  ]
                )
              case 1:
                return (f = r.sent()), [2, [s, df(f)]]
            }
          })
        })
      }),
      (t.URL_SCHEME_REGEX = /^https?:\/\//),
      t
    )
  })()
  function $f(t) {
    return null != t.match(Yf.URL_SCHEME_REGEX)
  }
  var Qf = function (t, e) {
    if ('undefined' == typeof fetch) return null
    return (
      Array.isArray(t)
        ? t.every(function (t) {
            return $f(t)
          })
        : $f(t)
    )
      ? Jf(t, { onProgress: e })
      : null
  }
  function Jf(t, e) {
    return new Yf(t, e)
  }
  mf.registerSaveRouter(Qf), mf.registerLoadRouter(Qf)
  var Zf = (function () {
      function t(t) {
        this.modelArtifacts = t
      }
      return (
        (t.prototype.load = function () {
          return n(this, void 0, void 0, function () {
            return r(this, function (t) {
              return [2, this.modelArtifacts]
            })
          })
        }),
        t
      )
    })(),
    td = (function () {
      function t(t) {
        this.saveHandler = t
      }
      return (
        (t.prototype.save = function (t) {
          return n(this, void 0, void 0, function () {
            return r(this, function (e) {
              return [2, this.saveHandler(t)]
            })
          })
        }),
        t
      )
    })()
  var ed = Object.freeze({
    browserFiles: function (t) {
      return new qf(t)
    },
    browserHTTPRequest: function (t, e) {
      return Jf(t, e)
    },
    concatenateArrayBuffers: df,
    decodeWeights: cf,
    encodeWeights: function (t, e) {
      return n(this, void 0, void 0, function () {
        var o,
          a,
          i,
          s,
          u,
          c = this
        return r(this, function (l) {
          switch (l.label) {
            case 0:
              for (
                o = [],
                  a = [],
                  i = Array.isArray(t)
                    ? t.map(function (t) {
                        return t.name
                      })
                    : Object.keys(t),
                  s = function (s) {
                    var u = i[s],
                      l = Array.isArray(t) ? t[s].tensor : t[u]
                    if (
                      'float32' !== l.dtype &&
                      'int32' !== l.dtype &&
                      'bool' !== l.dtype &&
                      'string' !== l.dtype
                    )
                      throw new Error(
                        "Unsupported dtype in weight '" + u + "': " + l.dtype
                      )
                    var h = { name: u, shape: l.shape, dtype: l.dtype }
                    if ('string' === l.dtype) {
                      var f = new Promise(function (t) {
                        return n(c, void 0, void 0, function () {
                          var e, n, o, a, i, s, u
                          return r(this, function (r) {
                            switch (r.label) {
                              case 0:
                                return [4, l.bytes()]
                              case 1:
                                for (
                                  e = r.sent(),
                                    n =
                                      e.reduce(function (t, e) {
                                        return t + e.length
                                      }, 0) +
                                      uf * e.length,
                                    o = new Uint8Array(n),
                                    a = 0,
                                    i = 0;
                                  i < e.length;
                                  i++
                                )
                                  (s = e[i]),
                                    (u = new Uint8Array(
                                      new Uint32Array([s.length]).buffer
                                    )),
                                    o.set(u, a),
                                    (a += uf),
                                    o.set(s, a),
                                    (a += s.length)
                                return t(o), [2]
                            }
                          })
                        })
                      })
                      a.push(f)
                    } else a.push(l.data())
                    null != e && (h.group = e), o.push(h)
                  },
                  u = 0;
                u < i.length;
                ++u
              )
                s(u)
              return [4, Promise.all(a)]
            case 1:
              return [2, { data: lf(l.sent()), specs: o }]
          }
        })
      })
    },
    fromMemory: function (t, e, n, r) {
      return 1 === arguments.length
        ? null != t.modelTopology || null != t.weightSpecs
          ? new Zf(t)
          : (console.warn(
              'Please call tf.io.fromMemory() with only one argument. The argument should be of type ModelArtifacts. The multi-argument signature of tf.io.fromMemory() has been deprecated and will be removed in a future release.'
            ),
            new Zf({ modelTopology: t }))
        : (console.warn(
            'Please call tf.io.fromMemory() with only one argument. The argument should be of type ModelArtifacts. The multi-argument signature of tf.io.fromMemory() has been deprecated and will be removed in a future release.'
          ),
          new Zf({
            modelTopology: t,
            weightSpecs: e,
            weightData: n,
            trainingConfig: r,
          }))
    },
    getLoadHandlers: function (t, e) {
      return mf.getLoadHandlers(t, e)
    },
    getModelArtifactsInfoForJSON: vf,
    getSaveHandlers: function (t) {
      return mf.getSaveHandlers(t)
    },
    http: Jf,
    isHTTPScheme: $f,
    loadWeights: function (t, e, o, a) {
      return (
        void 0 === e && (e = ''),
        n(this, void 0, void 0, function () {
          return r(this, function (n) {
            return [
              2,
              Xf(function (t) {
                return jf(t, { requestInit: a })
              })(t, e, o),
            ]
          })
        })
      )
    },
    registerLoadRouter: function (t) {
      return mf.registerLoadRouter(t)
    },
    registerSaveRouter: function (t) {
      return mf.registerSaveRouter(t)
    },
    weightsLoaderFactory: Xf,
    withSaveHandler: function (t) {
      return new td(t)
    },
    copyModel: function (t, e) {
      return n(this, void 0, void 0, function () {
        return r(this, function (n) {
          return [2, bf(t, e, !1)]
        })
      })
    },
    listModels: function () {
      return n(this, void 0, void 0, function () {
        var t, e, n, o, a, i, s
        return r(this, function (r) {
          switch (r.label) {
            case 0:
              ;(t = yf.getSchemes()), (e = {}), (n = 0), (o = t), (r.label = 1)
            case 1:
              return n < o.length
                ? ((a = o[n]), [4, yf.getManager(a).listModels()])
                : [3, 4]
            case 2:
              for (s in (i = r.sent())) e[a + gf + s] = i[s]
              r.label = 3
            case 3:
              return n++, [3, 1]
            case 4:
              return [2, e]
          }
        })
      })
    },
    moveModel: function (t, e) {
      return n(this, void 0, void 0, function () {
        return r(this, function (n) {
          return [2, bf(t, e, !0)]
        })
      })
    },
    removeModel: function (t) {
      return n(this, void 0, void 0, function () {
        var e
        return r(this, function (n) {
          return (e = xf(t)), [2, yf.getManager(e.scheme).removeModel(e.path)]
        })
      })
    },
  })
  var nd,
    rd = An({
      confusionMatrix_: function (t, e, n) {
        var r = mn(t, 'labels', 'confusionMatrix'),
          o = mn(e, 'predictions', 'confusionMatrix')
        C(null == n || (n > 0 && Number.isInteger(n)), function () {
          return (
            'If provided, numClasses must be a positive integer, but got ' + n
          )
        }),
          C(1 === r.rank, function () {
            return 'Expected the rank of labels to be 1, but got ' + r.rank
          }),
          C(1 === o.rank, function () {
            return 'Expected the rank of predictions to be 1, but got ' + o.rank
          }),
          C(r.shape[0] === o.shape[0], function () {
            return (
              'Mismatch in the number of examples: ' +
              r.shape[0] +
              ' vs. ' +
              o.shape[0] +
              '. Labels and predictions should have the same number of elements.'
            )
          }),
          C(n > 0 && Number.isInteger(n), function () {
            return (
              'numClasses is required to be a positive integer, but got ' + n
            )
          })
        var a = Rr(r.asType('int32'), n),
          i = Rr(o.asType('int32'), n)
        return a.transpose().matMul(i).asType('int32')
      },
    }),
    od = Object.freeze({ confusionMatrix: rd })
  var ad = An({
      fromPixels_: function (t, e) {
        if ((void 0 === e && (e = 3), e > 4))
          throw new Error(
            'Cannot construct Tensor with more than 4 channels from pixels.'
          )
        if (null == t)
          throw new Error(
            'pixels passed to tf.browser.fromPixels() can not be null'
          )
        var n = !1,
          r = !1,
          o = !1,
          a = !1,
          i = !1
        if (t.data instanceof Uint8Array) n = !0
        else if ('undefined' != typeof ImageData && t instanceof ImageData)
          r = !0
        else if (
          'undefined' != typeof HTMLVideoElement &&
          t instanceof HTMLVideoElement
        )
          o = !0
        else if (
          'undefined' != typeof HTMLImageElement &&
          t instanceof HTMLImageElement
        )
          a = !0
        else {
          if (null == t.getContext)
            throw new Error(
              'pixels passed to tf.browser.fromPixels() must be either an HTMLVideoElement, HTMLImageElement, HTMLCanvasElement, ImageData in browser, or OffscreenCanvas, ImageData in webworker or {data: Uint32Array, width: number, height: number}, but was ' +
                t.constructor.name
            )
          i = !0
        }
        if (o) {
          if (o && t.readyState < 2)
            throw new Error(
              'The video element has not loaded data yet. Please wait for `loadeddata` event on the <video> element.'
            )
        }
        if (null != l('FromPixels', Lt.backendName))
          return Lt.runKernel('FromPixels', { pixels: t }, { numChannels: e })
        var s,
          u,
          c = o ? [t.videoWidth, t.videoHeight] : [t.width, t.height],
          h = c[0],
          f = c[1]
        if (
          (i
            ? (s = t.getContext('2d').getImageData(0, 0, h, f).data)
            : r || n
            ? (s = t.data)
            : (a || o) &&
              (null == nd &&
                (nd = document.createElement('canvas').getContext('2d')),
              (nd.canvas.width = h),
              (nd.canvas.height = f),
              nd.drawImage(t, 0, 0, h, f),
              (s = nd.getImageData(0, 0, h, f).data)),
          4 === e)
        )
          u = new Int32Array(s)
        else {
          var d = h * f
          u = new Int32Array(d * e)
          for (var p = 0; p < d; p++)
            for (var v = 0; v < e; ++v) u[p * e + v] = s[4 * p + v]
        }
        return Pn(u, [f, h, e], 'int32')
      },
    }),
    id = Object.freeze({
      toPixels: function (t, e) {
        return n(this, void 0, void 0, function () {
          var n,
            o,
            a,
            i,
            s,
            u,
            c,
            l,
            h,
            f,
            d,
            p,
            v,
            m,
            g,
            y,
            x,
            b,
            w,
            C,
            E,
            R,
            I
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                if (
                  ((n = mn(t, 'img', 'toPixels')),
                  t instanceof wt || (n = n.toInt()),
                  2 !== n.rank && 3 !== n.rank)
                )
                  throw new Error(
                    'toPixels only supports rank 2 or 3 tensors, got rank ' +
                      n.rank +
                      '.'
                  )
                if (
                  ((o = n.shape.slice(0, 2)),
                  (a = o[0]),
                  (i = o[1]),
                  (s = 2 === n.rank ? 1 : n.shape[2]) > 4 || 2 === s)
                )
                  throw new Error(
                    'toPixels only supports depth of size 1, 3 or 4 but got ' +
                      s
                  )
                return [4, n.data()]
              case 1:
                return (
                  (u = r.sent()),
                  (c = n.min()),
                  (l = n.max()),
                  [4, Promise.all([c.data(), l.data()])]
                )
              case 2:
                if (
                  ((h = r.sent()),
                  (f = h[0]),
                  (d = h[1]),
                  (p = f[0]),
                  (v = d[0]),
                  c.dispose(),
                  l.dispose(),
                  'float32' === n.dtype)
                ) {
                  if (p < 0 || v > 1)
                    throw new Error(
                      'Tensor values for a float32 Tensor must be in the range [0 - 1] but got range [' +
                        p +
                        ' - ' +
                        v +
                        '].'
                    )
                } else {
                  if ('int32' !== n.dtype)
                    throw new Error(
                      'Unsupported type for toPixels: ' +
                        n.dtype +
                        '. Please use float32 or int32 tensors.'
                    )
                  if (p < 0 || v > 255)
                    throw new Error(
                      'Tensor values for a int32 Tensor must be in the range [0 - 255] but got range [' +
                        p +
                        ' - ' +
                        v +
                        '].'
                    )
                }
                for (
                  m = 'float32' === n.dtype ? 255 : 1,
                    g = new Uint8ClampedArray(i * a * 4),
                    y = 0;
                  y < a * i;
                  ++y
                )
                  (x = void 0),
                    (b = void 0),
                    (w = void 0),
                    (C = void 0),
                    1 === s
                      ? ((x = u[y] * m),
                        (b = u[y] * m),
                        (w = u[y] * m),
                        (C = 255))
                      : 3 === s
                      ? ((x = u[3 * y] * m),
                        (b = u[3 * y + 1] * m),
                        (w = u[3 * y + 2] * m),
                        (C = 255))
                      : 4 === s &&
                        ((x = u[4 * y] * m),
                        (b = u[4 * y + 1] * m),
                        (w = u[4 * y + 2] * m),
                        (C = u[4 * y + 3] * m)),
                    (g[(E = 4 * y) + 0] = Math.round(x)),
                    (g[E + 1] = Math.round(b)),
                    (g[E + 2] = Math.round(w)),
                    (g[E + 3] = Math.round(C))
                return (
                  null != e &&
                    ((e.width = i),
                    (e.height = a),
                    (R = e.getContext('2d')),
                    (I = new ImageData(g, i, a)),
                    R.putImageData(I, 0, 0)),
                  n !== t && n.dispose(),
                  [2, g]
                )
            }
          })
        })
      },
      fromPixels: ad,
    }),
    sd = (function () {
      function t() {}
      return (
        (t.prototype.getClassName = function () {
          return this.constructor.className
        }),
        (t.fromConfig = function (t, e) {
          return new t(e)
        }),
        t
      )
    })(),
    ud = (function () {
      function t() {
        this.classNameMap = {}
      }
      return (
        (t.getMap = function () {
          return null == t.instance && (t.instance = new t()), t.instance
        }),
        (t.register = function (e) {
          t.getMap().classNameMap[e.className] = [e, e.fromConfig]
        }),
        t
      )
    })()
  function cd(t) {
    C(null != t.className, function () {
      return 'Class being registered does not have the static className property defined.'
    }),
      C('string' == typeof t.className, function () {
        return (
          'className is required to be a string, but got type ' +
          typeof t.className
        )
      }),
      C(t.className.length > 0, function () {
        return 'Class being registered has an empty-string as its className, which is disallowed.'
      }),
      ud.register(t)
  }
  var ld = Object.freeze({
      Serializable: sd,
      SerializationMap: ud,
      registerClass: cd,
    }),
    hd = 0.001,
    fd = 0.1
  function dd() {
    return 32 === Lt.backend.floatPrecision() ? hd : fd
  }
  function pd(t, e, n) {
    var r = !0
    if (((V(t) || V(e)) && (r = !1), V(t) && V(e) && (r = !0), r)) {
      var o = t.constructor.name,
        a = e.constructor.name
      if (o !== a)
        throw new Error(
          'Arrays are of different type. Actual: ' + o + '. Expected: ' + a
        )
    }
    if (Array.isArray(t) && Array.isArray(e)) {
      var i = pn(t),
        s = pn(e)
      if (!S(i, s))
        throw new Error(
          'Arrays have different shapes. Actual: [' +
            i +
            ']. Expected: [' +
            s +
            ']'
        )
    }
    var u = V(t) ? t : I(t),
      c = V(e) ? e : I(e)
    if (u.length !== c.length)
      throw new Error(
        'Arrays have different lengths actual: ' +
          u.length +
          ' vs expected: ' +
          c.length +
          '.\nActual:   ' +
          u +
          '.\nExpected: ' +
          c +
          '.'
      )
    for (var l = 0; l < c.length; ++l) {
      var h = u[l],
        f = c[l]
      if (!n(h, f))
        throw new Error(
          'Arrays differ: actual[' +
            l +
            '] = ' +
            h +
            ', expected[' +
            l +
            '] = ' +
            f +
            '.\nActual:   ' +
            u +
            '.\nExpected: ' +
            c +
            '.'
        )
    }
  }
  function vd(t, e, n) {
    return (
      (!isFinite(t) && !isFinite(e)) ||
      !(isNaN(t) || isNaN(e) || Math.abs(t - e) > n)
    )
  }
  var md = Object.freeze({
      TEST_EPSILON_FLOAT16: fd,
      expectArraysClose: function (t, e, n) {
        return (
          null == n && (n = dd()),
          pd(t, e, function (t, e) {
            return vd(t, e, n)
          })
        )
      },
      testEpsilon: dd,
      expectPromiseToFail: function (t, e) {
        t().then(
          function () {
            return e.fail()
          },
          function () {
            return e()
          }
        )
      },
      expectArraysEqual: function (t, e) {
        var n =
          'string' == typeof e || 'number' == typeof e || 'boolean' == typeof e
            ? [e]
            : e
        return H(t) || H(t[0]) || H(e) || H(e[0])
          ? pd(t, n, function (t, e) {
              return t == e
            })
          : pd(t, e, function (t, e) {
              return vd(t, e, 0)
            })
      },
      expectNumbersClose: function (t, e, n) {
        if ((null == n && (n = dd()), !vd(t, e, n)))
          throw new Error(
            'Numbers differ: actual === ' + t + ', expected === ' + e
          )
      },
      expectValuesInRange: function (t, e, n) {
        for (var r = 0; r < t.length; r++)
          if (t[r] < e || t[r] > n)
            throw new Error(
              'Value out of range:' + t[r] + ' low: ' + e + ', high: ' + n
            )
      },
      expectArrayBuffersEqual: function (t, e) {
        expect(new Float32Array(t)).toEqual(new Float32Array(e))
      },
    }),
    gd = '1.7.0'
  var yd = Object.freeze({
      gpgpu_util: Mi,
      webgl_util: Ge,
      forceHalfFloat: function () {
        i().set('WEBGL_FORCE_F16_TEXTURES', !0)
      },
      MathBackendWebGL: Zs,
      setWebGLContext: Kt,
      GPGPUContext: Bi,
    }),
    xd = (function (t) {
      function o() {
        return (null !== t && t.apply(this, arguments)) || this
      }
      return (
        e(o, t),
        (o.prototype.minimize = function (t, e, n) {
          void 0 === e && (e = !1)
          var r = this.computeGradients(t, n),
            o = r.value,
            a = r.grads
          if (null != n) {
            var i = n.map(function (t) {
              return { name: t.name, tensor: a[t.name] }
            })
            this.applyGradients(i)
          } else this.applyGradients(a)
          return tn(a), e ? o : (o.dispose(), null)
        }),
        Object.defineProperty(o.prototype, 'iterations', {
          get: function () {
            return (
              null == this.iterations_ && (this.iterations_ = 0),
              this.iterations_
            )
          },
          enumerable: !0,
          configurable: !0,
        }),
        (o.prototype.incrementIterations = function () {
          this.iterations_ = this.iterations + 1
        }),
        (o.prototype.computeGradients = function (t, e) {
          return po(t, e)
        }),
        (o.prototype.dispose = function () {
          null != this.iterations_ && tn(this.iterations_)
        }),
        (o.prototype.saveIterations = function () {
          return n(this, void 0, void 0, function () {
            return r(this, function (t) {
              return (
                null == this.iterations_ && (this.iterations_ = 0),
                [2, { name: 'iter', tensor: On(this.iterations_, 'int32') }]
              )
            })
          })
        }),
        (o.prototype.getWeights = function () {
          return n(this, void 0, void 0, function () {
            return r(this, function (t) {
              throw new Error(
                'getWeights() is not implemented for this optimizer yet.'
              )
            })
          })
        }),
        (o.prototype.setWeights = function (t) {
          return n(this, void 0, void 0, function () {
            return r(this, function (t) {
              throw new Error(
                'setWeights() is not implemented for this optimizer class ' +
                  this.getClassName()
              )
            })
          })
        }),
        (o.prototype.extractIterations = function (t) {
          return n(this, void 0, void 0, function () {
            var e
            return r(this, function (n) {
              switch (n.label) {
                case 0:
                  return (e = this), [4, t[0].tensor.data()]
                case 1:
                  return (e.iterations_ = n.sent()[0]), [2, t.slice(1)]
              }
            })
          })
        }),
        o
      )
    })(sd)
  Object.defineProperty(xd, Symbol.hasInstance, {
    value: function (t) {
      return (
        null != t.minimize &&
        null != t.computeGradients &&
        null != t.applyGradients
      )
    },
  })
  var bd = (function (t) {
    function o(e, n, r) {
      void 0 === r && (r = null)
      var o = t.call(this) || this
      return (
        (o.learningRate = e),
        (o.rho = n),
        (o.epsilon = r),
        (o.accumulatedGrads = []),
        (o.accumulatedUpdates = []),
        null == r && (o.epsilon = Lt.backend.epsilon()),
        o
      )
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this
        ;(Array.isArray(t)
          ? t.map(function (t) {
              return t.name
            })
          : Object.keys(t)
        ).forEach(function (n, r) {
          var o = Lt.registeredVariables[n]
          null == e.accumulatedGrads[r] &&
            (e.accumulatedGrads[r] = {
              originalName: n + '/accum_grad',
              variable: Ze(function () {
                return Xn(o).variable(!1)
              }),
            }),
            null == e.accumulatedUpdates[r] &&
              (e.accumulatedUpdates[r] = {
                originalName: n + '/accum_var',
                variable: Ze(function () {
                  return Xn(o).variable(!1)
                }),
              })
          var a = Array.isArray(t) ? t[r].tensor : t[n]
          if (null != a) {
            var i = e.accumulatedGrads[r].variable,
              s = e.accumulatedUpdates[r].variable
            Ze(function () {
              var t = i.mul(e.rho).add(a.square().mul(1 - e.rho)),
                n = s.add(e.epsilon).sqrt().div(i.add(e.epsilon).sqrt()).mul(a),
                r = s.mul(e.rho).add(n.square().mul(1 - e.rho))
              i.assign(t), s.assign(r)
              var u = n.mul(-e.learningRate).add(o)
              o.assign(u)
            })
          }
        }),
          this.incrementIterations()
      }),
      (o.prototype.dispose = function () {
        null != this.accumulatedUpdates &&
          (tn(
            this.accumulatedGrads.map(function (t) {
              return t.variable
            })
          ),
          tn(
            this.accumulatedUpdates.map(function (t) {
              return t.variable
            })
          ))
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          var t
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return (
                  (t = this.accumulatedGrads.concat(this.accumulatedUpdates)),
                  [4, this.saveIterations()]
                )
              case 1:
                return [
                  2,
                  [e.sent()].concat(
                    t.map(function (t) {
                      return { name: t.originalName, tensor: t.variable }
                    })
                  ),
                ]
            }
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          var e
          return r(this, function (n) {
            switch (n.label) {
              case 0:
                return [4, this.extractIterations(t)]
              case 1:
                return (
                  (t = n.sent()),
                  (e = t.length / 2),
                  (this.accumulatedGrads = t.slice(0, e).map(function (t) {
                    return {
                      originalName: t.name,
                      variable: t.tensor.variable(!1),
                    }
                  })),
                  (this.accumulatedUpdates = t
                    .slice(e, 2 * e)
                    .map(function (t) {
                      return {
                        originalName: t.name,
                        variable: t.tensor.variable(!1),
                      }
                    })),
                  [2]
                )
            }
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return {
          learningRate: this.learningRate,
          rho: this.rho,
          epsilon: this.epsilon,
        }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate, e.rho, e.epsilon)
      }),
      (o.className = 'Adadelta'),
      o
    )
  })(xd)
  cd(bd)
  var wd = (function (t) {
    function o(e, n) {
      void 0 === n && (n = 0.1)
      var r = t.call(this) || this
      return (
        (r.learningRate = e),
        (r.initialAccumulatorValue = n),
        (r.accumulatedGrads = []),
        r
      )
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this
        ;(Array.isArray(t)
          ? t.map(function (t) {
              return t.name
            })
          : Object.keys(t)
        ).forEach(function (n, r) {
          var o = Lt.registeredVariables[n]
          if (null == e.accumulatedGrads[r]) {
            e.accumulatedGrads[r] = {
              originalName: n + '/accumulator',
              variable: Ze(function () {
                return Hn(o.shape, e.initialAccumulatorValue).variable(!1)
              }),
            }
          }
          var a = Array.isArray(t) ? t[r].tensor : t[n]
          if (null != a) {
            var i = e.accumulatedGrads[r].variable
            Ze(function () {
              var t = i.add(a.square())
              i.assign(t)
              var n = a
                .div(t.add(Lt.backend.epsilon()).sqrt())
                .mul(-e.learningRate)
                .add(o)
              o.assign(n)
            })
          }
        }),
          this.incrementIterations()
      }),
      (o.prototype.dispose = function () {
        null != this.accumulatedGrads &&
          tn(
            this.accumulatedGrads.map(function (t) {
              return t.variable
            })
          )
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          return r(this, function (t) {
            switch (t.label) {
              case 0:
                return [4, this.saveIterations()]
              case 1:
                return [
                  2,
                  [t.sent()].concat(
                    this.accumulatedGrads.map(function (t) {
                      return { name: t.originalName, tensor: t.variable }
                    })
                  ),
                ]
            }
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return [4, this.extractIterations(t)]
              case 1:
                return (
                  (t = e.sent()),
                  (this.accumulatedGrads = t.map(function (t) {
                    return {
                      originalName: t.name,
                      variable: t.tensor.variable(!1),
                    }
                  })),
                  [2]
                )
            }
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return {
          learningRate: this.learningRate,
          initialAccumulatorValue: this.initialAccumulatorValue,
        }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate, e.initialAccumulatorValue)
      }),
      (o.className = 'Adagrad'),
      o
    )
  })(xd)
  cd(wd)
  var Cd = (function (t) {
    function o(e, n, r, o) {
      void 0 === o && (o = null)
      var a = t.call(this) || this
      return (
        (a.learningRate = e),
        (a.beta1 = n),
        (a.beta2 = r),
        (a.epsilon = o),
        (a.accumulatedFirstMoment = []),
        (a.accumulatedSecondMoment = []),
        Ze(function () {
          ;(a.accBeta1 = On(n).variable()), (a.accBeta2 = On(r).variable())
        }),
        null == o && (a.epsilon = Lt.backend.epsilon()),
        a
      )
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this,
          n = Array.isArray(t)
            ? t.map(function (t) {
                return t.name
              })
            : Object.keys(t)
        Ze(function () {
          var r = Cc(1, e.accBeta1),
            o = Cc(1, e.accBeta2)
          n.forEach(function (n, a) {
            var i = Lt.registeredVariables[n]
            null == e.accumulatedFirstMoment[a] &&
              (e.accumulatedFirstMoment[a] = {
                originalName: n + '/m',
                variable: Ze(function () {
                  return Xn(i).variable(!1)
                }),
              }),
              null == e.accumulatedSecondMoment[a] &&
                (e.accumulatedSecondMoment[a] = {
                  originalName: n + '/v',
                  variable: Ze(function () {
                    return Xn(i).variable(!1)
                  }),
                })
            var s = Array.isArray(t) ? t[a].tensor : t[n]
            if (null != s) {
              var u = e.accumulatedFirstMoment[a].variable,
                c = e.accumulatedSecondMoment[a].variable,
                l = u.mul(e.beta1).add(s.mul(1 - e.beta1)),
                h = c.mul(e.beta2).add(s.square().mul(1 - e.beta2)),
                f = l.div(r),
                d = h.div(o)
              u.assign(l), c.assign(h)
              var p = f.div(d.sqrt().add(e.epsilon)).mul(-e.learningRate).add(i)
              i.assign(p)
            }
          }),
            e.accBeta1.assign(e.accBeta1.mul(e.beta1)),
            e.accBeta2.assign(e.accBeta2.mul(e.beta2))
        }),
          this.incrementIterations()
      }),
      (o.prototype.dispose = function () {
        this.accBeta1.dispose(),
          this.accBeta2.dispose(),
          null != this.accumulatedFirstMoment &&
            tn(
              this.accumulatedFirstMoment.map(function (t) {
                return t.variable
              })
            ),
          null != this.accumulatedSecondMoment &&
            tn(
              this.accumulatedSecondMoment.map(function (t) {
                return t.variable
              })
            )
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          var t
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return (
                  (t = this.accumulatedFirstMoment.concat(
                    this.accumulatedSecondMoment
                  )),
                  [4, this.saveIterations()]
                )
              case 1:
                return [
                  2,
                  [e.sent()].concat(
                    t.map(function (t) {
                      return { name: t.originalName, tensor: t.variable }
                    })
                  ),
                ]
            }
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          var e,
            n = this
          return r(this, function (r) {
            switch (r.label) {
              case 0:
                return [4, this.extractIterations(t)]
              case 1:
                return (
                  (t = r.sent()),
                  Ze(function () {
                    n.accBeta1.assign(xc(n.beta1, n.iterations_ + 1)),
                      n.accBeta2.assign(xc(n.beta2, n.iterations_ + 1))
                  }),
                  (e = t.length / 2),
                  (this.accumulatedFirstMoment = t
                    .slice(0, e)
                    .map(function (t) {
                      return {
                        originalName: t.name,
                        variable: t.tensor.variable(!1),
                      }
                    })),
                  (this.accumulatedSecondMoment = t
                    .slice(e, 2 * e)
                    .map(function (t) {
                      return {
                        originalName: t.name,
                        variable: t.tensor.variable(!1),
                      }
                    })),
                  [2]
                )
            }
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return {
          learningRate: this.learningRate,
          beta1: this.beta1,
          beta2: this.beta2,
          epsilon: this.epsilon,
        }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate, e.beta1, e.beta2, e.epsilon)
      }),
      (o.className = 'Adam'),
      o
    )
  })(xd)
  cd(Cd)
  var Ed = (function (t) {
    function o(e, n, r, o, a) {
      void 0 === o && (o = null), void 0 === a && (a = 0)
      var i = t.call(this) || this
      return (
        (i.learningRate = e),
        (i.beta1 = n),
        (i.beta2 = r),
        (i.epsilon = o),
        (i.decay = a),
        (i.accumulatedFirstMoment = []),
        (i.accumulatedWeightedInfNorm = []),
        Ze(function () {
          ;(i.iteration = On(0).variable()), (i.accBeta1 = On(n).variable())
        }),
        null == o && (i.epsilon = Lt.backend.epsilon()),
        i
      )
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this,
          n = Array.isArray(t)
            ? t.map(function (t) {
                return t.name
              })
            : Object.keys(t)
        Ze(function () {
          var r = Cc(1, e.accBeta1),
            o = sc(-e.learningRate, e.iteration.mul(e.decay).add(1))
          n.forEach(function (n, a) {
            var i = Lt.registeredVariables[n]
            null == e.accumulatedFirstMoment[a] &&
              (e.accumulatedFirstMoment[a] = {
                originalName: n + '/m',
                variable: Xn(i).variable(!1),
              }),
              null == e.accumulatedWeightedInfNorm[a] &&
                (e.accumulatedWeightedInfNorm[a] = {
                  originalName: n + '/v',
                  variable: Xn(i).variable(!1),
                })
            var s = Array.isArray(t) ? t[a].tensor : t[n]
            if (null != s) {
              var u = e.accumulatedFirstMoment[a].variable,
                c = e.accumulatedWeightedInfNorm[a].variable,
                l = u.mul(e.beta1).add(s.mul(1 - e.beta1)),
                h = c.mul(e.beta2),
                f = s.abs(),
                d = h.maximum(f)
              u.assign(l), c.assign(d)
              var p = o
                .div(r)
                .mul(l.div(d.add(e.epsilon)))
                .add(i)
              i.assign(p)
            }
          }),
            e.iteration.assign(e.iteration.add(1)),
            e.accBeta1.assign(e.accBeta1.mul(e.beta1))
        }),
          this.incrementIterations()
      }),
      (o.prototype.dispose = function () {
        this.accBeta1.dispose(),
          this.iteration.dispose(),
          null != this.accumulatedFirstMoment &&
            tn(
              this.accumulatedFirstMoment.map(function (t) {
                return t.variable
              })
            ),
          null != this.accumulatedWeightedInfNorm &&
            tn(
              this.accumulatedWeightedInfNorm.map(function (t) {
                return t.variable
              })
            )
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          return r(this, function (t) {
            throw new Error('getWeights() is not implemented for Adamax yet.')
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          return r(this, function (t) {
            throw new Error('setWeights() is not implemented for Adamax yet.')
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return {
          learningRate: this.learningRate,
          beta1: this.beta1,
          beta2: this.beta2,
          epsilon: this.epsilon,
          decay: this.decay,
        }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate, e.beta1, e.beta2, e.epsilon, e.decay)
      }),
      (o.className = 'Adamax'),
      o
    )
  })(xd)
  cd(Ed)
  var Rd = (function (t) {
    function o(e) {
      var n = t.call(this) || this
      return (n.learningRate = e), n.setLearningRate(e), n
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this
        ;(Array.isArray(t)
          ? t.map(function (t) {
              return t.name
            })
          : Object.keys(t)
        ).forEach(function (n, r) {
          var o = Array.isArray(t) ? t[r].tensor : t[n]
          if (null != o) {
            var a = Lt.registeredVariables[n]
            Ze(function () {
              var t = e.c.mul(o).add(a)
              a.assign(t)
            })
          }
        }),
          this.incrementIterations()
      }),
      (o.prototype.setLearningRate = function (t) {
        ;(this.learningRate = t),
          null != this.c && this.c.dispose(),
          (this.c = en(On(-t)))
      }),
      (o.prototype.dispose = function () {
        this.c.dispose()
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          return r(this, function (t) {
            switch (t.label) {
              case 0:
                return [4, this.saveIterations()]
              case 1:
                return [2, [t.sent()]]
            }
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return [4, this.extractIterations(t)]
              case 1:
                if (0 !== (t = e.sent()).length)
                  throw new Error(
                    'SGD optimizer does not have settable weights.'
                  )
                return [2]
            }
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return { learningRate: this.learningRate }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate)
      }),
      (o.className = 'SGD'),
      o
    )
  })(xd)
  cd(Rd)
  var Id = (function (t) {
    function o(e, n, r) {
      void 0 === r && (r = !1)
      var o = t.call(this, e) || this
      return (
        (o.learningRate = e),
        (o.momentum = n),
        (o.useNesterov = r),
        (o.accumulations = []),
        (o.m = On(o.momentum)),
        o
      )
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this
        ;(Array.isArray(t)
          ? t.map(function (t) {
              return t.name
            })
          : Object.keys(t)
        ).forEach(function (n, r) {
          var o = Lt.registeredVariables[n]
          if (null == e.accumulations[r]) {
            e.accumulations[r] = {
              originalName: n + '/momentum',
              variable: Ze(function () {
                return Xn(o).variable(!1)
              }),
            }
          }
          var a = e.accumulations[r].variable,
            i = Array.isArray(t) ? t[r].tensor : t[n]
          null != i &&
            Ze(function () {
              var t,
                n = e.m.mul(a).add(i)
              ;(t = e.useNesterov
                ? e.c.mul(i.add(n.mul(e.m))).add(o)
                : e.c.mul(n).add(o)),
                a.assign(n),
                o.assign(t)
            })
        }),
          this.incrementIterations()
      }),
      (o.prototype.dispose = function () {
        this.m.dispose(),
          null != this.accumulations &&
            tn(
              this.accumulations.map(function (t) {
                return t.variable
              })
            )
      }),
      (o.prototype.setMomentum = function (t) {
        this.momentum = t
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          return r(this, function (t) {
            switch (t.label) {
              case 0:
                return [4, this.saveIterations()]
              case 1:
                return [
                  2,
                  [t.sent()].concat(
                    this.accumulations.map(function (t) {
                      return { name: t.originalName, tensor: t.variable }
                    })
                  ),
                ]
            }
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return [4, this.extractIterations(t)]
              case 1:
                return (
                  (t = e.sent()),
                  (this.accumulations = t.map(function (t) {
                    return {
                      originalName: t.name,
                      variable: t.tensor.variable(!1),
                    }
                  })),
                  [2]
                )
            }
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return {
          learningRate: this.learningRate,
          momentum: this.momentum,
          useNesterov: this.useNesterov,
        }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate, e.momentum, e.useNesterov)
      }),
      (o.className = 'Momentum'),
      o
    )
  })(Rd)
  cd(Id)
  var kd = (function (t) {
    function o(e, n, r, o, a) {
      void 0 === n && (n = 0.9),
        void 0 === r && (r = 0),
        void 0 === o && (o = null),
        void 0 === a && (a = !1)
      var i = t.call(this) || this
      if (
        ((i.learningRate = e),
        (i.decay = n),
        (i.momentum = r),
        (i.epsilon = o),
        (i.accumulatedMeanSquares = []),
        (i.accumulatedMoments = []),
        (i.accumulatedMeanGrads = []),
        (i.centered = a),
        null == o && (i.epsilon = Lt.backend.epsilon()),
        null == e)
      )
        throw new Error('learningRate for RMSPropOptimizer must be defined.')
      return i
    }
    return (
      e(o, t),
      (o.prototype.applyGradients = function (t) {
        var e = this
        ;(Array.isArray(t)
          ? t.map(function (t) {
              return t.name
            })
          : Object.keys(t)
        ).forEach(function (n, r) {
          var o = Lt.registeredVariables[n]
          null == e.accumulatedMeanSquares[r] &&
            (e.accumulatedMeanSquares[r] = {
              originalName: n + '/rms',
              variable: Ze(function () {
                return Xn(o).variable(!1)
              }),
            }),
            null == e.accumulatedMoments[r] &&
              (e.accumulatedMoments[r] = {
                originalName: n + '/momentum',
                variable: Ze(function () {
                  return Xn(o).variable(!1)
                }),
              }),
            null == e.accumulatedMeanGrads[r] &&
              e.centered &&
              (e.accumulatedMeanGrads[r] = {
                originalName: n + '/mg',
                variable: Ze(function () {
                  return Xn(o).variable(!1)
                }),
              })
          var a = Array.isArray(t) ? t[r].tensor : t[n]
          if (null != a) {
            var i = e.accumulatedMeanSquares[r].variable,
              s = e.accumulatedMoments[r].variable
            Ze(function () {
              var t = i.mul(e.decay).add(a.square().mul(1 - e.decay))
              if (e.centered) {
                var n = e.accumulatedMeanGrads[r].variable,
                  u = n.mul(e.decay).add(a.mul(1 - e.decay)),
                  c = s
                    .mul(e.momentum)
                    .add(
                      a
                        .mul(e.learningRate)
                        .div(t.sub(u.square().add(e.epsilon)).sqrt())
                    )
                i.assign(t), n.assign(u), s.assign(c)
                var l = o.sub(c)
                o.assign(l)
              } else {
                var h = i.mul(e.decay).add(a.square().mul(1 - e.decay))
                c = s
                  .mul(e.momentum)
                  .add(a.mul(e.learningRate).div(h.add(e.epsilon).sqrt()))
                i.assign(h), s.assign(c)
                l = o.sub(c)
                o.assign(l)
              }
            })
          }
        }),
          this.incrementIterations()
      }),
      (o.prototype.dispose = function () {
        null != this.accumulatedMeanSquares &&
          tn(
            this.accumulatedMeanSquares.map(function (t) {
              return t.variable
            })
          ),
          null != this.accumulatedMeanGrads &&
            this.centered &&
            tn(
              this.accumulatedMeanGrads.map(function (t) {
                return t.variable
              })
            ),
          null != this.accumulatedMoments &&
            tn(
              this.accumulatedMoments.map(function (t) {
                return t.variable
              })
            )
      }),
      (o.prototype.getWeights = function () {
        return n(this, void 0, void 0, function () {
          var t
          return r(this, function (e) {
            switch (e.label) {
              case 0:
                return (
                  (t = this.accumulatedMeanSquares.concat(
                    this.accumulatedMoments
                  )),
                  this.centered && t.push.apply(t, this.accumulatedMeanGrads),
                  [4, this.saveIterations()]
                )
              case 1:
                return [
                  2,
                  [e.sent()].concat(
                    t.map(function (t) {
                      return { name: t.originalName, tensor: t.variable }
                    })
                  ),
                ]
            }
          })
        })
      }),
      (o.prototype.setWeights = function (t) {
        return n(this, void 0, void 0, function () {
          var e
          return r(this, function (n) {
            switch (n.label) {
              case 0:
                return [4, this.extractIterations(t)]
              case 1:
                return (
                  (t = n.sent()),
                  (e = this.centered ? t.length / 3 : t.length / 2),
                  (this.accumulatedMeanSquares = t
                    .slice(0, e)
                    .map(function (t) {
                      return {
                        originalName: t.name,
                        variable: t.tensor.variable(!1),
                      }
                    })),
                  (this.accumulatedMoments = t
                    .slice(e, 2 * e)
                    .map(function (t) {
                      return {
                        originalName: t.name,
                        variable: t.tensor.variable(!1),
                      }
                    })),
                  this.centered &&
                    (this.accumulatedMeanGrads = t
                      .slice(2 * e, 3 * e)
                      .map(function (t) {
                        return {
                          originalName: t.name,
                          variable: t.tensor.variable(!1),
                        }
                      })),
                  [2]
                )
            }
          })
        })
      }),
      (o.prototype.getConfig = function () {
        return {
          learningRate: this.learningRate,
          decay: this.decay,
          momentum: this.momentum,
          epsilon: this.epsilon,
          centered: this.centered,
        }
      }),
      (o.fromConfig = function (t, e) {
        return new t(e.learningRate, e.decay, e.momentum, e.epsilon, e.centered)
      }),
      (o.className = 'RMSProp'),
      o
    )
  })(xd)
  cd(kd)
  var Sd = (function () {
      function t() {}
      return (
        (t.sgd = function (t) {
          return new Rd(t)
        }),
        (t.momentum = function (t, e, n) {
          return void 0 === n && (n = !1), new Id(t, e, n)
        }),
        (t.rmsprop = function (t, e, n, r, o) {
          return (
            void 0 === e && (e = 0.9),
            void 0 === n && (n = 0),
            void 0 === r && (r = null),
            void 0 === o && (o = !1),
            new kd(t, e, n, r, o)
          )
        }),
        (t.adam = function (t, e, n, r) {
          return (
            void 0 === t && (t = 0.001),
            void 0 === e && (e = 0.9),
            void 0 === n && (n = 0.999),
            void 0 === r && (r = null),
            new Cd(t, e, n, r)
          )
        }),
        (t.adadelta = function (t, e, n) {
          return (
            void 0 === t && (t = 0.001),
            void 0 === e && (e = 0.95),
            void 0 === n && (n = null),
            new bd(t, e, n)
          )
        }),
        (t.adamax = function (t, e, n, r, o) {
          return (
            void 0 === t && (t = 0.002),
            void 0 === e && (e = 0.9),
            void 0 === n && (n = 0.999),
            void 0 === r && (r = null),
            void 0 === o && (o = 0),
            new Ed(t, e, n, r, o)
          )
        }),
        (t.adagrad = function (t, e) {
          return void 0 === e && (e = 0.1), new wd(t, e)
        }),
        t
      )
    })(),
    Ad = {
      sgd: Sd.sgd,
      momentum: Sd.momentum,
      adadelta: Sd.adadelta,
      adagrad: Sd.adagrad,
      rmsprop: Sd.rmsprop,
      adamax: Sd.adamax,
      adam: Sd.adam,
    },
    Dd =
      'undefined' != typeof requestAnimationFrame
        ? requestAnimationFrame
        : 'undefined' != typeof setImmediate
        ? setImmediate
        : function (t) {
            return t()
          }
  function Td() {
    return new Promise(function (t) {
      return Dd(function () {
        return t()
      })
    })
  }
  ;(wt.prototype.squaredDifference = function (t) {
    return nu(this, t)
  }),
    (xt = Gh)

  var tfCore_esm = /*#__PURE__*/ Object.freeze({
    __proto__: null,
    AdadeltaOptimizer: bd,
    AdagradOptimizer: wd,
    AdamOptimizer: Cd,
    AdamaxOptimizer: Ed,
    DataStorage: xo,
    get ENV() {
      return s
    },
    Environment: o,
    KernelBackend: bo,
    MomentumOptimizer: Id,
    Optimizer: xd,
    RMSPropOptimizer: kd,
    get Rank() {
      return Ct
    },
    get Reduction() {
      return ch
    },
    SGDOptimizer: Rd,
    Tensor: wt,
    TensorBuffer: gt,
    Variable: St,
    abs: ru,
    acos: ou,
    acosh: au,
    add: rc,
    addN: oc,
    addStrict: ac,
    all: Cl,
    any: El,
    argMax: Rl,
    argMin: Il,
    asin: iu,
    asinh: su,
    atan: uu,
    atan2: ic,
    atanh: cu,
    avgPool: fl,
    avgPool3d: vl,
    backend: hn,
    backend_util: Uo,
    basicLSTMCell: zl,
    batchNorm: ju,
    batchNorm2d: Xu,
    batchNorm3d: Yu,
    batchNorm4d: $u,
    batchNormalization: Ku,
    batchNormalization2d: Gu,
    batchNormalization3d: Hu,
    batchNormalization4d: qu,
    batchToSpaceND: vr,
    booleanMaskAsync: Uc,
    broadcastTo: mr,
    browser: id,
    buffer: dr,
    cast: gr,
    ceil: lu,
    clipByValue: hu,
    clone: yr,
    complex: Dn,
    concat: Yn,
    concat1d: $n,
    concat2d: Qn,
    concat3d: Jn,
    concat4d: Zn,
    conv1d: Hc,
    conv2d: qc,
    conv2dTranspose: Zc,
    conv3d: Kc,
    conv3dTranspose: tl,
    cos: fu,
    cosh: du,
    cumsum: xr,
    customGrad: vo,
    deprecationWarn: Xe,
    depthToSpace: br,
    depthwiseConv2d: Yc,
    diag: eh,
    disableDeprecationWarnings: je,
    dispose: tn,
    disposeVariables: Ye,
    div: sc,
    divNoNan: uc,
    divStrict: cc,
    dot: nl,
    dropout: nh,
    elu: _l,
    enableDebugMode: Ke,
    enableProdMode: qe,
    engine: $e,
    env: i,
    equal: Rc,
    equalStrict: Ic,
    erf: pu,
    exp: vu,
    expandDims: wr,
    expm1: mu,
    eye: Cr,
    fft: Xl,
    fill: Hn,
    findBackend: un,
    findBackendFactory: cn,
    floor: gu,
    floorDiv: lc,
    frame: ih,
    fused: zh,
    gather: Lc,
    gatherND: th,
    gather_util: Xr,
    getBackend: an,
    getGradient: h,
    getKernel: l,
    getKernelsForBackend: f,
    grad: co,
    grads: lo,
    greater: kc,
    greaterEqual: Sc,
    greaterEqualStrict: Ac,
    greaterStrict: Dc,
    hammingWindow: ah,
    hannWindow: oh,
    ifft: Yl,
    imag: Nn,
    image: Oh,
    inTopKAsync: lh,
    io: ed,
    irfft: Ql,
    isFinite: Du,
    isInf: Au,
    isNaN: Su,
    keep: en,
    leakyRelu: Ol,
    less: Tc,
    lessEqual: Nc,
    lessEqualStrict: Fc,
    lessStrict: _c,
    linalg: Ih,
    linspace: qn,
    localResponseNormalization: Ul,
    log: yu,
    log1p: xu,
    logSigmoid: bu,
    logSoftmax: yo,
    logSumExp: kl,
    logicalAnd: Qu,
    logicalNot: Ju,
    logicalOr: Zu,
    logicalXor: tc,
    losses: bh,
    matMul: el,
    math: od,
    max: Sl,
    maxPool: hl,
    maxPool3d: pl,
    maximum: hc,
    maximumStrict: fc,
    mean: Al,
    memory: Qe,
    min: Dl,
    minimum: dc,
    minimumStrict: pc,
    mod: vc,
    modStrict: mc,
    moments: Tl,
    movingAverage: Hl,
    mul: gc,
    mulStrict: yc,
    multiRNNCell: Gl,
    multinomial: Er,
    neg: wu,
    nextFrame: Td,
    norm: Vl,
    notEqual: Oc,
    notEqualStrict: Mc,
    oneHot: Rr,
    ones: zn,
    onesLike: jn,
    op: An,
    outerProduct: rl,
    pad: Ir,
    pad1d: kr,
    pad2d: Sr,
    pad3d: Ar,
    pad4d: Dr,
    pool: dl,
    pow: xc,
    powStrict: bc,
    prelu: Ml,
    print: pr,
    prod: Fl,
    profile: Je,
    rand: Tr,
    randomGamma: Fr,
    randomNormal: Nr,
    randomUniform: _r,
    range: Kn,
    ready: on,
    real: Tn,
    reciprocal: Cu,
    registerBackend: ln,
    registerGradient: p,
    registerKernel: d,
    relu: Bl,
    relu6: Pl,
    removeBackend: sn,
    reshape: Or,
    reverse: ol,
    reverse1d: al,
    reverse2d: il,
    reverse3d: sl,
    reverse4d: ul,
    rfft: $l,
    round: Eu,
    rsqrt: Ru,
    scalar: On,
    scatterND: jl,
    scatter_util: to,
    selu: Ll,
    separableConv2d: Jc,
    serialization: ld,
    setBackend: rn,
    setPlatform: fn,
    setdiff1dAsync: Vr,
    sigmoid: Iu,
    sign: ku,
    signal: uh,
    sin: Tu,
    sinh: Nu,
    slice: ml,
    slice1d: gl,
    slice2d: yl,
    slice3d: xl,
    slice4d: bl,
    slice_util: uo,
    softmax: go,
    softplus: Fu,
    spaceToBatchND: Mr,
    sparseToDense: Zl,
    spectral: Jl,
    split: tr,
    sqrt: _u,
    square: tu,
    squaredDifference: nu,
    squaredDifferenceStrict: wc,
    squeeze: Br,
    stack: Pr,
    step: Ou,
    stft: sh,
    stridedSlice: ql,
    sub: Cc,
    subStrict: Ec,
    sum: Nl,
    sumOutType: Tt,
    tan: Mu,
    tanh: Bu,
    tensor: Fn,
    tensor1d: Mn,
    tensor2d: Bn,
    tensor3d: Pn,
    tensor4d: Ln,
    tensor5d: Wn,
    tensor6d: Un,
    tensor_util: Mt,
    test_util: md,
    tidy: Ze,
    tile: Lr,
    time: nn,
    topk: Kl,
    train: Ad,
    transpose: Wl,
    truncatedNormal: Wr,
    unregisterGradient: m,
    unregisterKernel: v,
    unsortedSegmentSum: Wc,
    unstack: Ur,
    util: st,
    valueAndGrad: ho,
    valueAndGrads: fo,
    variable: Vn,
    variableGrads: po,
    version_core: gd,
    webgl: yd,
    where: ec,
    whereAsync: nc,
    zeros: Gn,
    zerosLike: Xn,
  })

  function drawContour(ctx, points, isClosed) {
    if (isClosed === void 0) {
      isClosed = false
    }
    ctx.beginPath()
    points.slice(1).forEach(function (_a, prevIdx) {
      var x = _a.x,
        y = _a.y
      var from = points[prevIdx]
      ctx.moveTo(from.x, from.y)
      ctx.lineTo(x, y)
    })
    if (isClosed) {
      var from = points[points.length - 1]
      var to = points[0]
      if (!from || !to) {
        return
      }
      ctx.moveTo(from.x, from.y)
      ctx.lineTo(to.x, to.y)
    }
    ctx.stroke()
  }

  /*! *****************************************************************************
    Copyright (c) Microsoft Corporation. All rights reserved.
    Licensed under the Apache License, Version 2.0 (the "License"); you may not use
    this file except in compliance with the License. You may obtain a copy of the
    License at http://www.apache.org/licenses/LICENSE-2.0
  
    THIS CODE IS PROVIDED ON AN *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
    KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
    WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
    MERCHANTABLITY OR NON-INFRINGEMENT.
  
    See the Apache Version 2.0 License for specific language governing permissions
    and limitations under the License.
    ***************************************************************************** */
  /* global Reflect, Promise */

  var extendStatics = function (d, b) {
    extendStatics =
      Object.setPrototypeOf ||
      ({ __proto__: [] } instanceof Array &&
        function (d, b) {
          d.__proto__ = b
        }) ||
      function (d, b) {
        for (var p in b) if (b.hasOwnProperty(p)) d[p] = b[p]
      }
    return extendStatics(d, b)
  }

  function __extends(d, b) {
    extendStatics(d, b)
    function __() {
      this.constructor = d
    }
    d.prototype =
      b === null ? Object.create(b) : ((__.prototype = b.prototype), new __())
  }

  var __assign = function () {
    __assign =
      Object.assign ||
      function __assign(t) {
        for (var s, i = 1, n = arguments.length; i < n; i++) {
          s = arguments[i]
          for (var p in s)
            if (Object.prototype.hasOwnProperty.call(s, p)) t[p] = s[p]
        }
        return t
      }
    return __assign.apply(this, arguments)
  }

  function __awaiter(thisArg, _arguments, P, generator) {
    return new (P || (P = Promise))(function (resolve, reject) {
      function fulfilled(value) {
        try {
          step(generator.next(value))
        } catch (e) {
          reject(e)
        }
      }
      function rejected(value) {
        try {
          step(generator['throw'](value))
        } catch (e) {
          reject(e)
        }
      }
      function step(result) {
        result.done
          ? resolve(result.value)
          : new P(function (resolve) {
              resolve(result.value)
            }).then(fulfilled, rejected)
      }
      step((generator = generator.apply(thisArg, _arguments || [])).next())
    })
  }

  function __generator(thisArg, body) {
    var _ = {
        label: 0,
        sent: function () {
          if (t[0] & 1) throw t[1]
          return t[1]
        },
        trys: [],
        ops: [],
      },
      f,
      y,
      t,
      g
    return (
      (g = { next: verb(0), throw: verb(1), return: verb(2) }),
      typeof Symbol === 'function' &&
        (g[Symbol.iterator] = function () {
          return this
        }),
      g
    )
    function verb(n) {
      return function (v) {
        return step([n, v])
      }
    }
    function step(op) {
      if (f) throw new TypeError('Generator is already executing.')
      while (_)
        try {
          if (
            ((f = 1),
            y &&
              (t =
                op[0] & 2
                  ? y['return']
                  : op[0]
                  ? y['throw'] || ((t = y['return']) && t.call(y), 0)
                  : y.next) &&
              !(t = t.call(y, op[1])).done)
          )
            return t
          if (((y = 0), t)) op = [op[0] & 2, t.value]
          switch (op[0]) {
            case 0:
            case 1:
              t = op
              break
            case 4:
              _.label++
              return { value: op[1], done: false }
            case 5:
              _.label++
              y = op[1]
              op = [0]
              continue
            case 7:
              op = _.ops.pop()
              _.trys.pop()
              continue
            default:
              if (
                !((t = _.trys), (t = t.length > 0 && t[t.length - 1])) &&
                (op[0] === 6 || op[0] === 2)
              ) {
                _ = 0
                continue
              }
              if (op[0] === 3 && (!t || (op[1] > t[0] && op[1] < t[3]))) {
                _.label = op[1]
                break
              }
              if (op[0] === 6 && _.label < t[1]) {
                _.label = t[1]
                t = op
                break
              }
              if (t && _.label < t[2]) {
                _.label = t[2]
                _.ops.push(op)
                break
              }
              if (t[2]) _.ops.pop()
              _.trys.pop()
              continue
          }
          op = body.call(thisArg, _)
        } catch (e) {
          op = [6, e]
          y = 0
        } finally {
          f = t = 0
        }
      if (op[0] & 5) throw op[1]
      return { value: op[0] ? op[1] : void 0, done: true }
    }
  }

  function __spreadArrays() {
    for (var s = 0, i = 0, il = arguments.length; i < il; i++)
      s += arguments[i].length
    for (var r = Array(s), k = 0, i = 0; i < il; i++)
      for (var a = arguments[i], j = 0, jl = a.length; j < jl; j++, k++)
        r[k] = a[j]
    return r
  }

  var Dimensions = /** @class */ (function () {
    function Dimensions(width, height) {
      if (!isValidNumber(width) || !isValidNumber(height)) {
        throw new Error(
          'Dimensions.constructor - expected width and height to be valid numbers, instead have ' +
            JSON.stringify({ width: width, height: height })
        )
      }
      this._width = width
      this._height = height
    }
    Object.defineProperty(Dimensions.prototype, 'width', {
      get: function () {
        return this._width
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Dimensions.prototype, 'height', {
      get: function () {
        return this._height
      },
      enumerable: true,
      configurable: true,
    })
    Dimensions.prototype.reverse = function () {
      return new Dimensions(1 / this.width, 1 / this.height)
    }
    return Dimensions
  })()

  function isTensor(tensor, dim) {
    return tensor instanceof wt && tensor.shape.length === dim
  }
  function isTensor1D(tensor) {
    return isTensor(tensor, 1)
  }
  function isTensor2D(tensor) {
    return isTensor(tensor, 2)
  }
  function isTensor3D(tensor) {
    return isTensor(tensor, 3)
  }
  function isTensor4D(tensor) {
    return isTensor(tensor, 4)
  }
  function isFloat(num) {
    return num % 1 !== 0
  }
  function isEven(num) {
    return num % 2 === 0
  }
  function round(num, prec) {
    if (prec === void 0) {
      prec = 2
    }
    var f = Math.pow(10, prec)
    return Math.floor(num * f) / f
  }
  function isDimensions(obj) {
    return obj && obj.width && obj.height
  }
  function computeReshapedDimensions(_a, inputSize) {
    var width = _a.width,
      height = _a.height
    var scale = inputSize / Math.max(height, width)
    return new Dimensions(Math.round(width * scale), Math.round(height * scale))
  }
  function getCenterPoint(pts) {
    return pts
      .reduce(function (sum, pt) {
        return sum.add(pt)
      }, new Point(0, 0))
      .div(new Point(pts.length, pts.length))
  }
  function range(num, start, step) {
    return Array(num)
      .fill(0)
      .map(function (_, i) {
        return start + i * step
      })
  }
  function isValidNumber(num) {
    return (
      (!!num && num !== Infinity && num !== -Infinity && !isNaN(num)) ||
      num === 0
    )
  }
  function isValidProbablitiy(num) {
    return isValidNumber(num) && 0 <= num && num <= 1.0
  }

  var index = /*#__PURE__*/ Object.freeze({
    __proto__: null,
    isTensor: isTensor,
    isTensor1D: isTensor1D,
    isTensor2D: isTensor2D,
    isTensor3D: isTensor3D,
    isTensor4D: isTensor4D,
    isFloat: isFloat,
    isEven: isEven,
    round: round,
    isDimensions: isDimensions,
    computeReshapedDimensions: computeReshapedDimensions,
    getCenterPoint: getCenterPoint,
    range: range,
    isValidNumber: isValidNumber,
    isValidProbablitiy: isValidProbablitiy,
  })

  var Point = /** @class */ (function () {
    function Point(x, y) {
      this._x = x
      this._y = y
    }
    Object.defineProperty(Point.prototype, 'x', {
      get: function () {
        return this._x
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Point.prototype, 'y', {
      get: function () {
        return this._y
      },
      enumerable: true,
      configurable: true,
    })
    Point.prototype.add = function (pt) {
      return new Point(this.x + pt.x, this.y + pt.y)
    }
    Point.prototype.sub = function (pt) {
      return new Point(this.x - pt.x, this.y - pt.y)
    }
    Point.prototype.mul = function (pt) {
      return new Point(this.x * pt.x, this.y * pt.y)
    }
    Point.prototype.div = function (pt) {
      return new Point(this.x / pt.x, this.y / pt.y)
    }
    Point.prototype.abs = function () {
      return new Point(Math.abs(this.x), Math.abs(this.y))
    }
    Point.prototype.magnitude = function () {
      return Math.sqrt(Math.pow(this.x, 2) + Math.pow(this.y, 2))
    }
    Point.prototype.floor = function () {
      return new Point(Math.floor(this.x), Math.floor(this.y))
    }
    return Point
  })()

  var Box = /** @class */ (function () {
    function Box(_box, allowNegativeDimensions) {
      if (allowNegativeDimensions === void 0) {
        allowNegativeDimensions = true
      }
      var box = _box || {}
      var isBbox = [box.left, box.top, box.right, box.bottom].every(
        isValidNumber
      )
      var isRect = [box.x, box.y, box.width, box.height].every(isValidNumber)
      if (!isRect && !isBbox) {
        throw new Error(
          'Box.constructor - expected box to be IBoundingBox | IRect, instead have ' +
            JSON.stringify(box)
        )
      }
      var _a = isRect
          ? [box.x, box.y, box.width, box.height]
          : [box.left, box.top, box.right - box.left, box.bottom - box.top],
        x = _a[0],
        y = _a[1],
        width = _a[2],
        height = _a[3]
      Box.assertIsValidBox(
        { x: x, y: y, width: width, height: height },
        'Box.constructor',
        allowNegativeDimensions
      )
      this._x = x
      this._y = y
      this._width = width
      this._height = height
    }
    Box.isRect = function (rect) {
      return (
        !!rect && [rect.x, rect.y, rect.width, rect.height].every(isValidNumber)
      )
    }
    Box.assertIsValidBox = function (box, callee, allowNegativeDimensions) {
      if (allowNegativeDimensions === void 0) {
        allowNegativeDimensions = false
      }
      if (!Box.isRect(box)) {
        throw new Error(
          callee +
            ' - invalid box: ' +
            JSON.stringify(box) +
            ', expected object with properties x, y, width, height'
        )
      }
      if (!allowNegativeDimensions && (box.width < 0 || box.height < 0)) {
        throw new Error(
          callee +
            ' - width (' +
            box.width +
            ') and height (' +
            box.height +
            ') must be positive numbers'
        )
      }
    }
    Object.defineProperty(Box.prototype, 'x', {
      get: function () {
        return this._x
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'y', {
      get: function () {
        return this._y
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'width', {
      get: function () {
        return this._width
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'height', {
      get: function () {
        return this._height
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'left', {
      get: function () {
        return this.x
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'top', {
      get: function () {
        return this.y
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'right', {
      get: function () {
        return this.x + this.width
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'bottom', {
      get: function () {
        return this.y + this.height
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'area', {
      get: function () {
        return this.width * this.height
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'topLeft', {
      get: function () {
        return new Point(this.left, this.top)
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'topRight', {
      get: function () {
        return new Point(this.right, this.top)
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'bottomLeft', {
      get: function () {
        return new Point(this.left, this.bottom)
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(Box.prototype, 'bottomRight', {
      get: function () {
        return new Point(this.right, this.bottom)
      },
      enumerable: true,
      configurable: true,
    })
    Box.prototype.round = function () {
      var _a = [this.x, this.y, this.width, this.height].map(function (val) {
          return Math.round(val)
        }),
        x = _a[0],
        y = _a[1],
        width = _a[2],
        height = _a[3]
      return new Box({ x: x, y: y, width: width, height: height })
    }
    Box.prototype.floor = function () {
      var _a = [this.x, this.y, this.width, this.height].map(function (val) {
          return Math.floor(val)
        }),
        x = _a[0],
        y = _a[1],
        width = _a[2],
        height = _a[3]
      return new Box({ x: x, y: y, width: width, height: height })
    }
    Box.prototype.toSquare = function () {
      var _a = this,
        x = _a.x,
        y = _a.y,
        width = _a.width,
        height = _a.height
      var diff = Math.abs(width - height)
      if (width < height) {
        x -= diff / 2
        width += diff
      }
      if (height < width) {
        y -= diff / 2
        height += diff
      }
      return new Box({ x: x, y: y, width: width, height: height })
    }
    Box.prototype.rescale = function (s) {
      var scaleX = isDimensions(s) ? s.width : s
      var scaleY = isDimensions(s) ? s.height : s
      return new Box({
        x: this.x * scaleX,
        y: this.y * scaleY,
        width: this.width * scaleX,
        height: this.height * scaleY,
      })
    }
    Box.prototype.pad = function (padX, padY) {
      var _a = [
          this.x - padX / 2,
          this.y - padY / 2,
          this.width + padX,
          this.height + padY,
        ],
        x = _a[0],
        y = _a[1],
        width = _a[2],
        height = _a[3]
      return new Box({ x: x, y: y, width: width, height: height })
    }
    Box.prototype.clipAtImageBorders = function (imgWidth, imgHeight) {
      var _a = this,
        x = _a.x,
        y = _a.y,
        right = _a.right,
        bottom = _a.bottom
      var clippedX = Math.max(x, 0)
      var clippedY = Math.max(y, 0)
      var newWidth = right - clippedX
      var newHeight = bottom - clippedY
      var clippedWidth = Math.min(newWidth, imgWidth - clippedX)
      var clippedHeight = Math.min(newHeight, imgHeight - clippedY)
      return new Box({
        x: clippedX,
        y: clippedY,
        width: clippedWidth,
        height: clippedHeight,
      }).floor()
    }
    Box.prototype.shift = function (sx, sy) {
      var _a = this,
        width = _a.width,
        height = _a.height
      var x = this.x + sx
      var y = this.y + sy
      return new Box({ x: x, y: y, width: width, height: height })
    }
    Box.prototype.padAtBorders = function (imageHeight, imageWidth) {
      var w = this.width + 1
      var h = this.height + 1
      var dx = 1
      var dy = 1
      var edx = w
      var edy = h
      var x = this.left
      var y = this.top
      var ex = this.right
      var ey = this.bottom
      if (ex > imageWidth) {
        edx = -ex + imageWidth + w
        ex = imageWidth
      }
      if (ey > imageHeight) {
        edy = -ey + imageHeight + h
        ey = imageHeight
      }
      if (x < 1) {
        edy = 2 - x
        x = 1
      }
      if (y < 1) {
        edy = 2 - y
        y = 1
      }
      return {
        dy: dy,
        edy: edy,
        dx: dx,
        edx: edx,
        y: y,
        ey: ey,
        x: x,
        ex: ex,
        w: w,
        h: h,
      }
    }
    Box.prototype.calibrate = function (region) {
      return new Box({
        left: this.left + region.left * this.width,
        top: this.top + region.top * this.height,
        right: this.right + region.right * this.width,
        bottom: this.bottom + region.bottom * this.height,
      })
        .toSquare()
        .round()
    }
    return Box
  })()

  var BoundingBox = /** @class */ (function (_super) {
    __extends(BoundingBox, _super)
    function BoundingBox(left, top, right, bottom, allowNegativeDimensions) {
      if (allowNegativeDimensions === void 0) {
        allowNegativeDimensions = false
      }
      return (
        _super.call(
          this,
          { left: left, top: top, right: right, bottom: bottom },
          allowNegativeDimensions
        ) || this
      )
    }
    return BoundingBox
  })(Box)

  var ObjectDetection = /** @class */ (function () {
    function ObjectDetection(
      score,
      classScore,
      className,
      relativeBox,
      imageDims
    ) {
      this._imageDims = new Dimensions(imageDims.width, imageDims.height)
      this._score = score
      this._classScore = classScore
      this._className = className
      this._box = new Box(relativeBox).rescale(this._imageDims)
    }
    Object.defineProperty(ObjectDetection.prototype, 'score', {
      get: function () {
        return this._score
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'classScore', {
      get: function () {
        return this._classScore
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'className', {
      get: function () {
        return this._className
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'box', {
      get: function () {
        return this._box
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'imageDims', {
      get: function () {
        return this._imageDims
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'imageWidth', {
      get: function () {
        return this.imageDims.width
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'imageHeight', {
      get: function () {
        return this.imageDims.height
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(ObjectDetection.prototype, 'relativeBox', {
      get: function () {
        return new Box(this._box).rescale(this.imageDims.reverse())
      },
      enumerable: true,
      configurable: true,
    })
    ObjectDetection.prototype.forSize = function (width, height) {
      return new ObjectDetection(
        this.score,
        this.classScore,
        this.className,
        this.relativeBox,
        { width: width, height: height }
      )
    }
    return ObjectDetection
  })()

  var FaceDetection = /** @class */ (function (_super) {
    __extends(FaceDetection, _super)
    function FaceDetection(score, relativeBox, imageDims) {
      return _super.call(this, score, score, '', relativeBox, imageDims) || this
    }
    FaceDetection.prototype.forSize = function (width, height) {
      var _a = _super.prototype.forSize.call(this, width, height),
        score = _a.score,
        relativeBox = _a.relativeBox,
        imageDims = _a.imageDims
      return new FaceDetection(score, relativeBox, imageDims)
    }
    return FaceDetection
  })(ObjectDetection)

  function iou(box1, box2, isIOU) {
    if (isIOU === void 0) {
      isIOU = true
    }
    var width = Math.max(
      0.0,
      Math.min(box1.right, box2.right) - Math.max(box1.left, box2.left)
    )
    var height = Math.max(
      0.0,
      Math.min(box1.bottom, box2.bottom) - Math.max(box1.top, box2.top)
    )
    var interSection = width * height
    return isIOU
      ? interSection / (box1.area + box2.area - interSection)
      : interSection / Math.min(box1.area, box2.area)
  }

  function minBbox(pts) {
    var xs = pts.map(function (pt) {
      return pt.x
    })
    var ys = pts.map(function (pt) {
      return pt.y
    })
    var minX = xs.reduce(function (min, x) {
      return x < min ? x : min
    }, Infinity)
    var minY = ys.reduce(function (min, y) {
      return y < min ? y : min
    }, Infinity)
    var maxX = xs.reduce(function (max, x) {
      return max < x ? x : max
    }, 0)
    var maxY = ys.reduce(function (max, y) {
      return max < y ? y : max
    }, 0)
    return new BoundingBox(minX, minY, maxX, maxY)
  }

  function nonMaxSuppression(boxes, scores, iouThreshold, isIOU) {
    if (isIOU === void 0) {
      isIOU = true
    }
    var indicesSortedByScore = scores
      .map(function (score, boxIndex) {
        return { score: score, boxIndex: boxIndex }
      })
      .sort(function (c1, c2) {
        return c1.score - c2.score
      })
      .map(function (c) {
        return c.boxIndex
      })
    var pick = []
    var _loop_1 = function () {
      var curr = indicesSortedByScore.pop()
      pick.push(curr)
      var indices = indicesSortedByScore
      var outputs = []
      for (var i = 0; i < indices.length; i++) {
        var idx = indices[i]
        var currBox = boxes[curr]
        var idxBox = boxes[idx]
        outputs.push(iou(currBox, idxBox, isIOU))
      }
      indicesSortedByScore = indicesSortedByScore.filter(function (_, j) {
        return outputs[j] <= iouThreshold
      })
    }
    while (indicesSortedByScore.length > 0) {
      _loop_1()
    }
    return pick
  }

  function normalize(x, meanRgb) {
    return Ze(function () {
      var r = meanRgb[0],
        g = meanRgb[1],
        b = meanRgb[2]
      var avg_r = Hn(__spreadArrays(x.shape.slice(0, 3), [1]), r)
      var avg_g = Hn(__spreadArrays(x.shape.slice(0, 3), [1]), g)
      var avg_b = Hn(__spreadArrays(x.shape.slice(0, 3), [1]), b)
      var avg_rgb = Yn([avg_r, avg_g, avg_b], 3)
      return Cc(x, avg_rgb)
    })
  }

  /**
   * Pads the smaller dimension of an image tensor with zeros, such that width === height.
   *
   * @param imgTensor The image tensor.
   * @param isCenterImage (optional, default: false) If true, add an equal amount of padding on
   * both sides of the minor dimension oof the image.
   * @returns The padded tensor with width === height.
   */
  function padToSquare(imgTensor, isCenterImage) {
    if (isCenterImage === void 0) {
      isCenterImage = false
    }
    return Ze(function () {
      var _a = imgTensor.shape.slice(1),
        height = _a[0],
        width = _a[1]
      if (height === width) {
        return imgTensor
      }
      var dimDiff = Math.abs(height - width)
      var paddingAmount = Math.round(dimDiff * (isCenterImage ? 0.5 : 1))
      var paddingAxis = height > width ? 2 : 1
      var createPaddingTensor = function (paddingAmount) {
        var paddingTensorShape = imgTensor.shape.slice()
        paddingTensorShape[paddingAxis] = paddingAmount
        return Hn(paddingTensorShape, 0)
      }
      var paddingTensorAppend = createPaddingTensor(paddingAmount)
      var remainingPaddingAmount =
        dimDiff - paddingTensorAppend.shape[paddingAxis]
      var paddingTensorPrepend =
        isCenterImage && remainingPaddingAmount
          ? createPaddingTensor(remainingPaddingAmount)
          : null
      var tensorsToStack = [
        paddingTensorPrepend,
        imgTensor,
        paddingTensorAppend,
      ]
        .filter(function (t) {
          return !!t
        })
        .map(function (t) {
          return t.toFloat()
        })
      return Yn(tensorsToStack, paddingAxis)
    })
  }

  function shuffleArray(inputArray) {
    var array = inputArray.slice()
    for (var i = array.length - 1; i > 0; i--) {
      var j = Math.floor(Math.random() * (i + 1))
      var x = array[i]
      array[i] = array[j]
      array[j] = x
    }
    return array
  }

  function sigmoid(x) {
    return 1 / (1 + Math.exp(-x))
  }
  function inverseSigmoid(x) {
    return Math.log(x / (1 - x))
  }

  var Rect = /** @class */ (function (_super) {
    __extends(Rect, _super)
    function Rect(x, y, width, height, allowNegativeDimensions) {
      if (allowNegativeDimensions === void 0) {
        allowNegativeDimensions = false
      }
      return (
        _super.call(
          this,
          { x: x, y: y, width: width, height: height },
          allowNegativeDimensions
        ) || this
      )
    }
    return Rect
  })(Box)

  // face alignment constants
  var relX = 0.5
  var relY = 0.43
  var relScale = 0.45
  var FaceLandmarks = /** @class */ (function () {
    function FaceLandmarks(relativeFaceLandmarkPositions, imgDims, shift) {
      if (shift === void 0) {
        shift = new Point(0, 0)
      }
      var width = imgDims.width,
        height = imgDims.height
      this._imgDims = new Dimensions(width, height)
      this._shift = shift
      this._positions = relativeFaceLandmarkPositions.map(function (pt) {
        return pt.mul(new Point(width, height)).add(shift)
      })
    }
    Object.defineProperty(FaceLandmarks.prototype, 'shift', {
      get: function () {
        return new Point(this._shift.x, this._shift.y)
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(FaceLandmarks.prototype, 'imageWidth', {
      get: function () {
        return this._imgDims.width
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(FaceLandmarks.prototype, 'imageHeight', {
      get: function () {
        return this._imgDims.height
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(FaceLandmarks.prototype, 'positions', {
      get: function () {
        return this._positions
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(FaceLandmarks.prototype, 'relativePositions', {
      get: function () {
        var _this = this
        return this._positions.map(function (pt) {
          return pt
            .sub(_this._shift)
            .div(new Point(_this.imageWidth, _this.imageHeight))
        })
      },
      enumerable: true,
      configurable: true,
    })
    FaceLandmarks.prototype.forSize = function (width, height) {
      return new this.constructor(this.relativePositions, {
        width: width,
        height: height,
      })
    }
    FaceLandmarks.prototype.shiftBy = function (x, y) {
      return new this.constructor(
        this.relativePositions,
        this._imgDims,
        new Point(x, y)
      )
    }
    FaceLandmarks.prototype.shiftByPoint = function (pt) {
      return this.shiftBy(pt.x, pt.y)
    }
    /**
     * Aligns the face landmarks after face detection from the relative positions of the faces
     * bounding box, or it's current shift. This function should be used to align the face images
     * after face detection has been performed, before they are passed to the face recognition net.
     * This will make the computed face descriptor more accurate.
     *
     * @param detection (optional) The bounding box of the face or the face detection result. If
     * no argument was passed the position of the face landmarks are assumed to be relative to
     * it's current shift.
     * @returns The bounding box of the aligned face.
     */
    FaceLandmarks.prototype.align = function (detection, options) {
      if (options === void 0) {
        options = {}
      }
      if (detection) {
        var box =
          detection instanceof FaceDetection
            ? detection.box.floor()
            : new Box(detection)
        return this.shiftBy(box.x, box.y).align(null, options)
      }
      var _a = Object.assign(
          {},
          { useDlibAlignment: false, minBoxPadding: 0.2 },
          options
        ),
        useDlibAlignment = _a.useDlibAlignment,
        minBoxPadding = _a.minBoxPadding
      if (useDlibAlignment) {
        return this.alignDlib()
      }
      return this.alignMinBbox(minBoxPadding)
    }
    FaceLandmarks.prototype.alignDlib = function () {
      var centers = this.getRefPointsForAlignment()
      var leftEyeCenter = centers[0],
        rightEyeCenter = centers[1],
        mouthCenter = centers[2]
      var distToMouth = function (pt) {
        return mouthCenter.sub(pt).magnitude()
      }
      var eyeToMouthDist =
        (distToMouth(leftEyeCenter) + distToMouth(rightEyeCenter)) / 2
      var size = Math.floor(eyeToMouthDist / relScale)
      var refPoint = getCenterPoint(centers)
      // TODO: pad in case rectangle is out of image bounds
      var x = Math.floor(Math.max(0, refPoint.x - relX * size))
      var y = Math.floor(Math.max(0, refPoint.y - relY * size))
      return new Rect(
        x,
        y,
        Math.min(size, this.imageWidth + x),
        Math.min(size, this.imageHeight + y)
      )
    }
    FaceLandmarks.prototype.alignMinBbox = function (padding) {
      var box = minBbox(this.positions)
      return box.pad(box.width * padding, box.height * padding)
    }
    FaceLandmarks.prototype.getRefPointsForAlignment = function () {
      throw new Error('getRefPointsForAlignment not implemented by base class')
    }
    return FaceLandmarks
  })()

  var FaceLandmarks5 = /** @class */ (function (_super) {
    __extends(FaceLandmarks5, _super)
    function FaceLandmarks5() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    FaceLandmarks5.prototype.getRefPointsForAlignment = function () {
      var pts = this.positions
      return [pts[0], pts[1], getCenterPoint([pts[3], pts[4]])]
    }
    return FaceLandmarks5
  })(FaceLandmarks)

  var FaceLandmarks68 = /** @class */ (function (_super) {
    __extends(FaceLandmarks68, _super)
    function FaceLandmarks68() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    FaceLandmarks68.prototype.getJawOutline = function () {
      return this.positions.slice(0, 17)
    }
    FaceLandmarks68.prototype.getLeftEyeBrow = function () {
      return this.positions.slice(17, 22)
    }
    FaceLandmarks68.prototype.getRightEyeBrow = function () {
      return this.positions.slice(22, 27)
    }
    FaceLandmarks68.prototype.getNose = function () {
      return this.positions.slice(27, 36)
    }
    FaceLandmarks68.prototype.getLeftEye = function () {
      return this.positions.slice(36, 42)
    }
    FaceLandmarks68.prototype.getRightEye = function () {
      return this.positions.slice(42, 48)
    }
    FaceLandmarks68.prototype.getMouth = function () {
      return this.positions.slice(48, 68)
    }
    FaceLandmarks68.prototype.getRefPointsForAlignment = function () {
      return [this.getLeftEye(), this.getRightEye(), this.getMouth()].map(
        getCenterPoint
      )
    }
    return FaceLandmarks68
  })(FaceLandmarks)

  var FaceMatch = /** @class */ (function () {
    function FaceMatch(label, distance) {
      this._label = label
      this._distance = distance
    }
    Object.defineProperty(FaceMatch.prototype, 'label', {
      get: function () {
        return this._label
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(FaceMatch.prototype, 'distance', {
      get: function () {
        return this._distance
      },
      enumerable: true,
      configurable: true,
    })
    FaceMatch.prototype.toString = function (withDistance) {
      if (withDistance === void 0) {
        withDistance = true
      }
      return (
        '' +
        this.label +
        (withDistance ? ' (' + round(this.distance) + ')' : '')
      )
    }
    return FaceMatch
  })()

  var LabeledBox = /** @class */ (function (_super) {
    __extends(LabeledBox, _super)
    function LabeledBox(box, label) {
      var _this = _super.call(this, box) || this
      _this._label = label
      return _this
    }
    LabeledBox.assertIsValidLabeledBox = function (box, callee) {
      Box.assertIsValidBox(box, callee)
      if (!isValidNumber(box.label)) {
        throw new Error(
          callee +
            ' - expected property label (' +
            box.label +
            ') to be a number'
        )
      }
    }
    Object.defineProperty(LabeledBox.prototype, 'label', {
      get: function () {
        return this._label
      },
      enumerable: true,
      configurable: true,
    })
    return LabeledBox
  })(Box)

  var LabeledFaceDescriptors = /** @class */ (function () {
    function LabeledFaceDescriptors(label, descriptors) {
      if (!(typeof label === 'string')) {
        throw new Error(
          'LabeledFaceDescriptors - constructor expected label to be a string'
        )
      }
      if (
        !Array.isArray(descriptors) ||
        descriptors.some(function (desc) {
          return !(desc instanceof Float32Array)
        })
      ) {
        throw new Error(
          'LabeledFaceDescriptors - constructor expected descriptors to be an array of Float32Array'
        )
      }
      this._label = label
      this._descriptors = descriptors
    }
    Object.defineProperty(LabeledFaceDescriptors.prototype, 'label', {
      get: function () {
        return this._label
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(LabeledFaceDescriptors.prototype, 'descriptors', {
      get: function () {
        return this._descriptors
      },
      enumerable: true,
      configurable: true,
    })
    LabeledFaceDescriptors.prototype.toJSON = function () {
      return {
        label: this.label,
        descriptors: this.descriptors.map(function (d) {
          return Array.from(d)
        }),
      }
    }
    LabeledFaceDescriptors.fromJSON = function (json) {
      var descriptors = json.descriptors.map(function (d) {
        return new Float32Array(d)
      })
      return new LabeledFaceDescriptors(json.label, descriptors)
    }
    return LabeledFaceDescriptors
  })()

  var PredictedBox = /** @class */ (function (_super) {
    __extends(PredictedBox, _super)
    function PredictedBox(box, label, score, classScore) {
      var _this = _super.call(this, box, label) || this
      _this._score = score
      _this._classScore = classScore
      return _this
    }
    PredictedBox.assertIsValidPredictedBox = function (box, callee) {
      LabeledBox.assertIsValidLabeledBox(box, callee)
      if (
        !isValidProbablitiy(box.score) ||
        !isValidProbablitiy(box.classScore)
      ) {
        throw new Error(
          callee +
            ' - expected properties score (' +
            box.score +
            ') and (' +
            box.classScore +
            ') to be a number between [0, 1]'
        )
      }
    }
    Object.defineProperty(PredictedBox.prototype, 'score', {
      get: function () {
        return this._score
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(PredictedBox.prototype, 'classScore', {
      get: function () {
        return this._classScore
      },
      enumerable: true,
      configurable: true,
    })
    return PredictedBox
  })(LabeledBox)

  function isWithFaceDetection(obj) {
    return obj['detection'] instanceof FaceDetection
  }
  function extendWithFaceDetection(sourceObj, detection) {
    var extension = { detection: detection }
    return Object.assign({}, sourceObj, extension)
  }

  function createBrowserEnv() {
    var fetch =
      window['fetch'] ||
      function () {
        throw new Error(
          'fetch - missing fetch implementation for browser environment'
        )
      }
    var readFile = function () {
      throw new Error(
        'readFile - filesystem not available for browser environment'
      )
    }
    return {
      Canvas: HTMLCanvasElement,
      CanvasRenderingContext2D: CanvasRenderingContext2D,
      Image: HTMLImageElement,
      ImageData: ImageData,
      Video: HTMLVideoElement,
      createCanvasElement: function () {
        return document.createElement('canvas')
      },
      createImageElement: function () {
        return document.createElement('img')
      },
      fetch: fetch,
      readFile: readFile,
    }
  }

  function createFileSystem(fs) {
    var requireFsError = ''
    if (!fs) {
      try {
        fs = require('fs')
      } catch (err) {
        requireFsError = err.toString()
      }
    }
    var readFile = fs
      ? function (filePath) {
          return new Promise(function (res, rej) {
            fs.readFile(filePath, function (err, buffer) {
              return err ? rej(err) : res(buffer)
            })
          })
        }
      : function () {
          throw new Error(
            'readFile - failed to require fs in nodejs environment with error: ' +
              requireFsError
          )
        }
    return {
      readFile: readFile,
    }
  }

  function createNodejsEnv() {
    var Canvas = global['Canvas'] || global['HTMLCanvasElement']
    var Image = global['Image'] || global['HTMLImageElement']
    var createCanvasElement = function () {
      if (Canvas) {
        return new Canvas()
      }
      throw new Error(
        'createCanvasElement - missing Canvas implementation for nodejs environment'
      )
    }
    var createImageElement = function () {
      if (Image) {
        return new Image()
      }
      throw new Error(
        'createImageElement - missing Image implementation for nodejs environment'
      )
    }
    var fetch =
      global['fetch'] ||
      function () {
        throw new Error(
          'fetch - missing fetch implementation for nodejs environment'
        )
      }
    var fileSystem = createFileSystem()
    return __assign(
      {
        Canvas:
          Canvas ||
          /** @class */ (function () {
            function Canvas() {}
            return Canvas
          })(),
        CanvasRenderingContext2D:
          global['CanvasRenderingContext2D'] ||
          /** @class */ (function () {
            function class_1() {}
            return class_1
          })(),
        Image:
          Image ||
          /** @class */ (function () {
            function Image() {}
            return Image
          })(),
        ImageData:
          global['ImageData'] ||
          /** @class */ (function () {
            function class_2() {}
            return class_2
          })(),
        Video:
          global['HTMLVideoElement'] ||
          /** @class */ (function () {
            function class_3() {}
            return class_3
          })(),
        createCanvasElement: createCanvasElement,
        createImageElement: createImageElement,
        fetch: fetch,
      },
      fileSystem
    )
  }

  function isBrowser() {
    return (
      typeof window === 'object' &&
      typeof document !== 'undefined' &&
      typeof HTMLImageElement !== 'undefined' &&
      typeof HTMLCanvasElement !== 'undefined' &&
      typeof HTMLVideoElement !== 'undefined' &&
      typeof ImageData !== 'undefined' &&
      typeof CanvasRenderingContext2D !== 'undefined'
    )
  }

  function isNodejs() {
    return (
      typeof global === 'object' &&
      typeof require === 'function' &&
      typeof module !== 'undefined' &&
      // issues with gatsby.js: module.exports is undefined
      // && !!module.exports
      typeof process !== 'undefined' &&
      !!process.version
    )
  }

  var environment
  function getEnv() {
    if (!environment) {
      throw new Error(
        'getEnv - environment is not defined, check isNodejs() and isBrowser()'
      )
    }
    return environment
  }
  function setEnv(env) {
    environment = env
  }
  function initialize() {
    // check for isBrowser() first to prevent electron renderer process
    // to be initialized with wrong environment due to isNodejs() returning true
    if (isBrowser()) {
      setEnv(createBrowserEnv())
    }
    if (isNodejs()) {
      setEnv(createNodejsEnv())
    }
  }
  function monkeyPatch(env) {
    if (!environment) {
      initialize()
    }
    if (!environment) {
      throw new Error(
        'monkeyPatch - environment is not defined, check isNodejs() and isBrowser()'
      )
    }
    var _a = env.Canvas,
      Canvas = _a === void 0 ? environment.Canvas : _a,
      _b = env.Image,
      Image = _b === void 0 ? environment.Image : _b
    environment.Canvas = Canvas
    environment.Image = Image
    environment.createCanvasElement =
      env.createCanvasElement ||
      function () {
        return new Canvas()
      }
    environment.createImageElement =
      env.createImageElement ||
      function () {
        return new Image()
      }
    environment.ImageData = env.ImageData || environment.ImageData
    environment.Video = env.Video || environment.Video
    environment.fetch = env.fetch || environment.fetch
    environment.readFile = env.readFile || environment.readFile
  }
  var env = {
    getEnv: getEnv,
    setEnv: setEnv,
    initialize: initialize,
    createBrowserEnv: createBrowserEnv,
    createFileSystem: createFileSystem,
    createNodejsEnv: createNodejsEnv,
    monkeyPatch: monkeyPatch,
    isBrowser: isBrowser,
    isNodejs: isNodejs,
  }
  initialize()

  function resolveInput(arg) {
    if (!env.isNodejs() && typeof arg === 'string') {
      return document.getElementById(arg)
    }
    return arg
  }

  function getContext2dOrThrow(canvasArg) {
    var _a = env.getEnv(),
      Canvas = _a.Canvas,
      CanvasRenderingContext2D = _a.CanvasRenderingContext2D
    if (canvasArg instanceof CanvasRenderingContext2D) {
      return canvasArg
    }
    var canvas = resolveInput(canvasArg)
    if (!(canvas instanceof Canvas)) {
      throw new Error(
        'resolveContext2d - expected canvas to be of instance of Canvas'
      )
    }
    var ctx = canvas.getContext('2d')
    if (!ctx) {
      throw new Error('resolveContext2d - canvas 2d context is null')
    }
    return ctx
  }

  var AnchorPosition
  ;(function (AnchorPosition) {
    AnchorPosition['TOP_LEFT'] = 'TOP_LEFT'
    AnchorPosition['TOP_RIGHT'] = 'TOP_RIGHT'
    AnchorPosition['BOTTOM_LEFT'] = 'BOTTOM_LEFT'
    AnchorPosition['BOTTOM_RIGHT'] = 'BOTTOM_RIGHT'
  })(AnchorPosition || (AnchorPosition = {}))
  var DrawTextFieldOptions = /** @class */ (function () {
    function DrawTextFieldOptions(options) {
      if (options === void 0) {
        options = {}
      }
      var anchorPosition = options.anchorPosition,
        backgroundColor = options.backgroundColor,
        fontColor = options.fontColor,
        fontSize = options.fontSize,
        fontStyle = options.fontStyle,
        padding = options.padding
      this.anchorPosition = anchorPosition || AnchorPosition.TOP_LEFT
      this.backgroundColor = backgroundColor || 'rgba(0, 0, 0, 0.5)'
      this.fontColor = fontColor || 'rgba(255, 255, 255, 1)'
      this.fontSize = fontSize || 14
      this.fontStyle = fontStyle || 'Georgia'
      this.padding = padding || 4
    }
    return DrawTextFieldOptions
  })()
  var DrawTextField = /** @class */ (function () {
    function DrawTextField(text, anchor, options) {
      if (options === void 0) {
        options = {}
      }
      this.text =
        typeof text === 'string'
          ? [text]
          : text instanceof DrawTextField
          ? text.text
          : text
      this.anchor = anchor
      this.options = new DrawTextFieldOptions(options)
    }
    DrawTextField.prototype.measureWidth = function (ctx) {
      var padding = this.options.padding
      return (
        this.text
          .map(function (l) {
            return ctx.measureText(l).width
          })
          .reduce(function (w0, w1) {
            return w0 < w1 ? w1 : w0
          }, 0) +
        2 * padding
      )
    }
    DrawTextField.prototype.measureHeight = function () {
      var _a = this.options,
        fontSize = _a.fontSize,
        padding = _a.padding
      return this.text.length * fontSize + 2 * padding
    }
    DrawTextField.prototype.getUpperLeft = function (ctx, canvasDims) {
      var anchorPosition = this.options.anchorPosition
      var isShiftLeft =
        anchorPosition === AnchorPosition.BOTTOM_RIGHT ||
        anchorPosition === AnchorPosition.TOP_RIGHT
      var isShiftTop =
        anchorPosition === AnchorPosition.BOTTOM_LEFT ||
        anchorPosition === AnchorPosition.BOTTOM_RIGHT
      var textFieldWidth = this.measureWidth(ctx)
      var textFieldHeight = this.measureHeight()
      var x = isShiftLeft ? this.anchor.x - textFieldWidth : this.anchor.x
      var y = isShiftTop ? this.anchor.y - textFieldHeight : this.anchor.y
      // adjust anchor if text box exceeds canvas borders
      if (canvasDims) {
        var width = canvasDims.width,
          height = canvasDims.height
        var newX = Math.max(Math.min(x, width - textFieldWidth), 0)
        var newY = Math.max(Math.min(y, height - textFieldHeight), 0)
        return { x: newX, y: newY }
      }
      return { x: x, y: y }
    }
    DrawTextField.prototype.draw = function (canvasArg) {
      var canvas = resolveInput(canvasArg)
      var ctx = getContext2dOrThrow(canvas)
      var _a = this.options,
        backgroundColor = _a.backgroundColor,
        fontColor = _a.fontColor,
        fontSize = _a.fontSize,
        fontStyle = _a.fontStyle,
        padding = _a.padding
      ctx.font = fontSize + 'px ' + fontStyle
      var maxTextWidth = this.measureWidth(ctx)
      var textHeight = this.measureHeight()
      ctx.fillStyle = backgroundColor
      var upperLeft = this.getUpperLeft(ctx, canvas)
      ctx.fillRect(upperLeft.x, upperLeft.y, maxTextWidth, textHeight)
      ctx.fillStyle = fontColor
      this.text.forEach(function (textLine, i) {
        var x = padding + upperLeft.x
        var y = padding + upperLeft.y + (i + 1) * fontSize
        ctx.fillText(textLine, x, y)
      })
    }
    return DrawTextField
  })()

  var DrawBoxOptions = /** @class */ (function () {
    function DrawBoxOptions(options) {
      if (options === void 0) {
        options = {}
      }
      var boxColor = options.boxColor,
        lineWidth = options.lineWidth,
        label = options.label,
        drawLabelOptions = options.drawLabelOptions
      this.boxColor = boxColor || 'rgba(0, 0, 255, 1)'
      this.lineWidth = lineWidth || 2
      this.label = label
      var defaultDrawLabelOptions = {
        anchorPosition: AnchorPosition.BOTTOM_LEFT,
        backgroundColor: this.boxColor,
      }
      this.drawLabelOptions = new DrawTextFieldOptions(
        Object.assign({}, defaultDrawLabelOptions, drawLabelOptions)
      )
    }
    return DrawBoxOptions
  })()
  var DrawBox = /** @class */ (function () {
    function DrawBox(box, options) {
      if (options === void 0) {
        options = {}
      }
      this.box = new Box(box)
      this.options = new DrawBoxOptions(options)
    }
    DrawBox.prototype.draw = function (canvasArg) {
      var ctx = getContext2dOrThrow(canvasArg)
      var _a = this.options,
        boxColor = _a.boxColor,
        lineWidth = _a.lineWidth
      var _b = this.box,
        x = _b.x,
        y = _b.y,
        width = _b.width,
        height = _b.height
      ctx.strokeStyle = boxColor
      ctx.lineWidth = lineWidth
      ctx.strokeRect(x, y, width, height)
      var label = this.options.label
      if (label) {
        new DrawTextField(
          [label],
          { x: x - lineWidth / 2, y: y },
          this.options.drawLabelOptions
        ).draw(canvasArg)
      }
    }
    return DrawBox
  })()

  function drawDetections(canvasArg, detections) {
    var detectionsArray = Array.isArray(detections) ? detections : [detections]
    detectionsArray.forEach(function (det) {
      var score =
        det instanceof FaceDetection
          ? det.score
          : isWithFaceDetection(det)
          ? det.detection.score
          : undefined
      var box =
        det instanceof FaceDetection
          ? det.box
          : isWithFaceDetection(det)
          ? det.detection.box
          : new Box(det)
      var label = score ? '' + round(score) : undefined
      new DrawBox(box, { label: label }).draw(canvasArg)
    })
  }

  function isMediaLoaded(media) {
    var _a = env.getEnv(),
      Image = _a.Image,
      Video = _a.Video
    return (
      (media instanceof Image && media.complete) ||
      (media instanceof Video && media.readyState >= 3)
    )
  }

  function awaitMediaLoaded(media) {
    return new Promise(function (resolve, reject) {
      if (media instanceof env.getEnv().Canvas || isMediaLoaded(media)) {
        return resolve()
      }
      function onLoad(e) {
        if (!e.currentTarget) return
        e.currentTarget.removeEventListener('load', onLoad)
        e.currentTarget.removeEventListener('error', onError)
        resolve(e)
      }
      function onError(e) {
        if (!e.currentTarget) return
        e.currentTarget.removeEventListener('load', onLoad)
        e.currentTarget.removeEventListener('error', onError)
        reject(e)
      }
      media.addEventListener('load', onLoad)
      media.addEventListener('error', onError)
    })
  }

  function bufferToImage(buf) {
    return new Promise(function (resolve, reject) {
      if (!(buf instanceof Blob)) {
        return reject('bufferToImage - expected buf to be of type: Blob')
      }
      var reader = new FileReader()
      reader.onload = function () {
        if (typeof reader.result !== 'string') {
          return reject(
            'bufferToImage - expected reader.result to be a string, in onload'
          )
        }
        var img = env.getEnv().createImageElement()
        img.onload = function () {
          return resolve(img)
        }
        img.onerror = reject
        img.src = reader.result
      }
      reader.onerror = reject
      reader.readAsDataURL(buf)
    })
  }

  function getMediaDimensions(input) {
    var _a = env.getEnv(),
      Image = _a.Image,
      Video = _a.Video
    if (input instanceof Image) {
      return new Dimensions(input.naturalWidth, input.naturalHeight)
    }
    if (input instanceof Video) {
      return new Dimensions(input.videoWidth, input.videoHeight)
    }
    return new Dimensions(input.width, input.height)
  }

  function createCanvas(_a) {
    var width = _a.width,
      height = _a.height
    var createCanvasElement = env.getEnv().createCanvasElement
    var canvas = createCanvasElement()
    canvas.width = width
    canvas.height = height
    return canvas
  }
  function createCanvasFromMedia(media, dims) {
    var ImageData = env.getEnv().ImageData
    if (!(media instanceof ImageData) && !isMediaLoaded(media)) {
      throw new Error(
        'createCanvasFromMedia - media has not finished loading yet'
      )
    }
    var _a = dims || getMediaDimensions(media),
      width = _a.width,
      height = _a.height
    var canvas = createCanvas({ width: width, height: height })
    if (media instanceof ImageData) {
      getContext2dOrThrow(canvas).putImageData(media, 0, 0)
    } else {
      getContext2dOrThrow(canvas).drawImage(media, 0, 0, width, height)
    }
    return canvas
  }

  function imageTensorToCanvas(imgTensor, canvas) {
    return __awaiter(this, void 0, void 0, function () {
      var targetCanvas, _a, height, width, numChannels, imgTensor3D
      return __generator(this, function (_b) {
        switch (_b.label) {
          case 0:
            targetCanvas = canvas || env.getEnv().createCanvasElement()
            ;(_a = imgTensor.shape.slice(isTensor4D(imgTensor) ? 1 : 0)),
              (height = _a[0]),
              (width = _a[1]),
              (numChannels = _a[2])
            imgTensor3D = Ze(function () {
              return imgTensor.as3D(height, width, numChannels).toInt()
            })
            return [4 /*yield*/, id.toPixels(imgTensor3D, targetCanvas)]
          case 1:
            _b.sent()
            imgTensor3D.dispose()
            return [2 /*return*/, targetCanvas]
        }
      })
    })
  }

  function isMediaElement(input) {
    var _a = env.getEnv(),
      Image = _a.Image,
      Canvas = _a.Canvas,
      Video = _a.Video
    return (
      input instanceof Image ||
      input instanceof Canvas ||
      input instanceof Video
    )
  }

  function imageToSquare(input, inputSize, centerImage) {
    if (centerImage === void 0) {
      centerImage = false
    }
    var _a = env.getEnv(),
      Image = _a.Image,
      Canvas = _a.Canvas
    if (!(input instanceof Image || input instanceof Canvas)) {
      throw new Error(
        'imageToSquare - expected arg0 to be HTMLImageElement | HTMLCanvasElement'
      )
    }
    var dims = getMediaDimensions(input)
    var scale = inputSize / Math.max(dims.height, dims.width)
    var width = scale * dims.width
    var height = scale * dims.height
    var targetCanvas = createCanvas({ width: inputSize, height: inputSize })
    var inputCanvas =
      input instanceof Canvas ? input : createCanvasFromMedia(input)
    var offset = Math.abs(width - height) / 2
    var dx = centerImage && width < height ? offset : 0
    var dy = centerImage && height < width ? offset : 0
    getContext2dOrThrow(targetCanvas).drawImage(
      inputCanvas,
      dx,
      dy,
      width,
      height
    )
    return targetCanvas
  }

  var NetInput = /** @class */ (function () {
    function NetInput(inputs, treatAsBatchInput) {
      var _this = this
      if (treatAsBatchInput === void 0) {
        treatAsBatchInput = false
      }
      this._imageTensors = []
      this._canvases = []
      this._treatAsBatchInput = false
      this._inputDimensions = []
      if (!Array.isArray(inputs)) {
        throw new Error(
          'NetInput.constructor - expected inputs to be an Array of TResolvedNetInput or to be instanceof tf.Tensor4D, instead have ' +
            inputs
        )
      }
      this._treatAsBatchInput = treatAsBatchInput
      this._batchSize = inputs.length
      inputs.forEach(function (input, idx) {
        if (isTensor3D(input)) {
          _this._imageTensors[idx] = input
          _this._inputDimensions[idx] = input.shape
          return
        }
        if (isTensor4D(input)) {
          var batchSize = input.shape[0]
          if (batchSize !== 1) {
            throw new Error(
              'NetInput - tf.Tensor4D with batchSize ' +
                batchSize +
                ' passed, but not supported in input array'
            )
          }
          _this._imageTensors[idx] = input
          _this._inputDimensions[idx] = input.shape.slice(1)
          return
        }
        var canvas =
          input instanceof env.getEnv().Canvas
            ? input
            : createCanvasFromMedia(input)
        _this._canvases[idx] = canvas
        _this._inputDimensions[idx] = [canvas.height, canvas.width, 3]
      })
    }
    Object.defineProperty(NetInput.prototype, 'imageTensors', {
      get: function () {
        return this._imageTensors
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NetInput.prototype, 'canvases', {
      get: function () {
        return this._canvases
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NetInput.prototype, 'isBatchInput', {
      get: function () {
        return this.batchSize > 1 || this._treatAsBatchInput
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NetInput.prototype, 'batchSize', {
      get: function () {
        return this._batchSize
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NetInput.prototype, 'inputDimensions', {
      get: function () {
        return this._inputDimensions
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NetInput.prototype, 'inputSize', {
      get: function () {
        return this._inputSize
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NetInput.prototype, 'reshapedInputDimensions', {
      get: function () {
        var _this = this
        return range(this.batchSize, 0, 1).map(function (_, batchIdx) {
          return _this.getReshapedInputDimensions(batchIdx)
        })
      },
      enumerable: true,
      configurable: true,
    })
    NetInput.prototype.getInput = function (batchIdx) {
      return this.canvases[batchIdx] || this.imageTensors[batchIdx]
    }
    NetInput.prototype.getInputDimensions = function (batchIdx) {
      return this._inputDimensions[batchIdx]
    }
    NetInput.prototype.getInputHeight = function (batchIdx) {
      return this._inputDimensions[batchIdx][0]
    }
    NetInput.prototype.getInputWidth = function (batchIdx) {
      return this._inputDimensions[batchIdx][1]
    }
    NetInput.prototype.getReshapedInputDimensions = function (batchIdx) {
      if (typeof this.inputSize !== 'number') {
        throw new Error(
          'getReshapedInputDimensions - inputSize not set, toBatchTensor has not been called yet'
        )
      }
      var width = this.getInputWidth(batchIdx)
      var height = this.getInputHeight(batchIdx)
      return computeReshapedDimensions(
        { width: width, height: height },
        this.inputSize
      )
    }
    /**
     * Create a batch tensor from all input canvases and tensors
     * with size [batchSize, inputSize, inputSize, 3].
     *
     * @param inputSize Height and width of the tensor.
     * @param isCenterImage (optional, default: false) If true, add an equal amount of padding on
     * both sides of the minor dimension oof the image.
     * @returns The batch tensor.
     */
    NetInput.prototype.toBatchTensor = function (inputSize, isCenterInputs) {
      var _this = this
      if (isCenterInputs === void 0) {
        isCenterInputs = true
      }
      this._inputSize = inputSize
      return Ze(function () {
        var inputTensors = range(_this.batchSize, 0, 1).map(function (
          batchIdx
        ) {
          var input = _this.getInput(batchIdx)
          if (input instanceof wt) {
            var imgTensor = isTensor4D(input) ? input : input.expandDims()
            imgTensor = padToSquare(imgTensor, isCenterInputs)
            if (
              imgTensor.shape[1] !== inputSize ||
              imgTensor.shape[2] !== inputSize
            ) {
              imgTensor = Oh.resizeBilinear(imgTensor, [inputSize, inputSize])
            }
            return imgTensor.as3D(inputSize, inputSize, 3)
          }
          if (input instanceof env.getEnv().Canvas) {
            return id.fromPixels(
              imageToSquare(input, inputSize, isCenterInputs)
            )
          }
          throw new Error(
            'toBatchTensor - at batchIdx ' +
              batchIdx +
              ', expected input to be instanceof tf.Tensor or instanceof HTMLCanvasElement, instead have ' +
              input
          )
        })
        var batchTensor = Pr(
          inputTensors.map(function (t) {
            return t.toFloat()
          })
        ).as4D(_this.batchSize, inputSize, inputSize, 3)
        return batchTensor
      })
    }
    return NetInput
  })()

  /**
   * Validates the input to make sure, they are valid net inputs and awaits all media elements
   * to be finished loading.
   *
   * @param input The input, which can be a media element or an array of different media elements.
   * @returns A NetInput instance, which can be passed into one of the neural networks.
   */
  function toNetInput(inputs) {
    return __awaiter(this, void 0, void 0, function () {
      var inputArgArray, getIdxHint, inputArray
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            if (inputs instanceof NetInput) {
              return [2 /*return*/, inputs]
            }
            inputArgArray = Array.isArray(inputs) ? inputs : [inputs]
            if (!inputArgArray.length) {
              throw new Error('toNetInput - empty array passed as input')
            }
            getIdxHint = function (idx) {
              return Array.isArray(inputs) ? ' at input index ' + idx + ':' : ''
            }
            inputArray = inputArgArray.map(resolveInput)
            inputArray.forEach(function (input, i) {
              if (
                !isMediaElement(input) &&
                !isTensor3D(input) &&
                !isTensor4D(input)
              ) {
                if (typeof inputArgArray[i] === 'string') {
                  throw new Error(
                    'toNetInput -' +
                      getIdxHint(i) +
                      ' string passed, but could not resolve HTMLElement for element id ' +
                      inputArgArray[i]
                  )
                }
                throw new Error(
                  'toNetInput -' +
                    getIdxHint(i) +
                    ' expected media to be of type HTMLImageElement | HTMLVideoElement | HTMLCanvasElement | tf.Tensor3D, or to be an element id'
                )
              }
              if (isTensor4D(input)) {
                // if tf.Tensor4D is passed in the input array, the batch size has to be 1
                var batchSize = input.shape[0]
                if (batchSize !== 1) {
                  throw new Error(
                    'toNetInput -' +
                      getIdxHint(i) +
                      ' tf.Tensor4D with batchSize ' +
                      batchSize +
                      ' passed, but not supported in input array'
                  )
                }
              }
            })
            // wait for all media elements being loaded
            return [
              4 /*yield*/,
              Promise.all(
                inputArray.map(function (input) {
                  return isMediaElement(input) && awaitMediaLoaded(input)
                })
              ),
            ]
          case 1:
            // wait for all media elements being loaded
            _a.sent()
            return [
              2 /*return*/,
              new NetInput(inputArray, Array.isArray(inputs)),
            ]
        }
      })
    })
  }

  /**
   * Extracts the image regions containing the detected faces.
   *
   * @param input The image that face detection has been performed on.
   * @param detections The face detection results or face bounding boxes for that image.
   * @returns The Canvases of the corresponding image region for each detected face.
   */
  function extractFaces(input, detections) {
    return __awaiter(this, void 0, void 0, function () {
      var Canvas, canvas, netInput, tensorOrCanvas, _a, ctx, boxes
      return __generator(this, function (_b) {
        switch (_b.label) {
          case 0:
            Canvas = env.getEnv().Canvas
            canvas = input
            if (!!(input instanceof Canvas)) return [3 /*break*/, 5]
            return [4 /*yield*/, toNetInput(input)]
          case 1:
            netInput = _b.sent()
            if (netInput.batchSize > 1) {
              throw new Error('extractFaces - batchSize > 1 not supported')
            }
            tensorOrCanvas = netInput.getInput(0)
            if (!(tensorOrCanvas instanceof Canvas)) return [3 /*break*/, 2]
            _a = tensorOrCanvas
            return [3 /*break*/, 4]
          case 2:
            return [4 /*yield*/, imageTensorToCanvas(tensorOrCanvas)]
          case 3:
            _a = _b.sent()
            _b.label = 4
          case 4:
            canvas = _a
            _b.label = 5
          case 5:
            ctx = getContext2dOrThrow(canvas)
            boxes = detections
              .map(function (det) {
                return det instanceof FaceDetection
                  ? det.forSize(canvas.width, canvas.height).box.floor()
                  : det
              })
              .map(function (box) {
                return box.clipAtImageBorders(canvas.width, canvas.height)
              })
            return [
              2 /*return*/,
              boxes.map(function (_a) {
                var x = _a.x,
                  y = _a.y,
                  width = _a.width,
                  height = _a.height
                var faceImg = createCanvas({ width: width, height: height })
                getContext2dOrThrow(faceImg).putImageData(
                  ctx.getImageData(x, y, width, height),
                  0,
                  0
                )
                return faceImg
              }),
            ]
        }
      })
    })
  }

  /**
   * Extracts the tensors of the image regions containing the detected faces.
   * Useful if you want to compute the face descriptors for the face images.
   * Using this method is faster then extracting a canvas for each face and
   * converting them to tensors individually.
   *
   * @param imageTensor The image tensor that face detection has been performed on.
   * @param detections The face detection results or face bounding boxes for that image.
   * @returns Tensors of the corresponding image region for each detected face.
   */
  function extractFaceTensors(imageTensor, detections) {
    return __awaiter(this, void 0, void 0, function () {
      return __generator(this, function (_a) {
        if (!isTensor3D(imageTensor) && !isTensor4D(imageTensor)) {
          throw new Error(
            'extractFaceTensors - expected image tensor to be 3D or 4D'
          )
        }
        if (isTensor4D(imageTensor) && imageTensor.shape[0] > 1) {
          throw new Error('extractFaceTensors - batchSize > 1 not supported')
        }
        return [
          2 /*return*/,
          Ze(function () {
            var _a = imageTensor.shape.slice(isTensor4D(imageTensor) ? 1 : 0),
              imgHeight = _a[0],
              imgWidth = _a[1],
              numChannels = _a[2]
            var boxes = detections
              .map(function (det) {
                return det instanceof FaceDetection
                  ? det.forSize(imgWidth, imgHeight).box
                  : det
              })
              .map(function (box) {
                return box.clipAtImageBorders(imgWidth, imgHeight)
              })
            var faceTensors = boxes.map(function (_a) {
              var x = _a.x,
                y = _a.y,
                width = _a.width,
                height = _a.height
              return xl(
                imageTensor.as3D(imgHeight, imgWidth, numChannels),
                [y, x, 0],
                [height, width, numChannels]
              )
            })
            return faceTensors
          }),
        ]
      })
    })
  }

  function fetchOrThrow(url, init) {
    return __awaiter(this, void 0, void 0, function () {
      var fetch, res
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            fetch = env.getEnv().fetch
            return [4 /*yield*/, fetch(url, init)]
          case 1:
            res = _a.sent()
            if (!(res.status < 400)) {
              throw new Error(
                'failed to fetch: (' +
                  res.status +
                  ') ' +
                  res.statusText +
                  ', from url: ' +
                  res.url
              )
            }
            return [2 /*return*/, res]
        }
      })
    })
  }

  function fetchImage(uri) {
    return __awaiter(this, void 0, void 0, function () {
      var res, blob
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            return [4 /*yield*/, fetchOrThrow(uri)]
          case 1:
            res = _a.sent()
            return [4 /*yield*/, res.blob()]
          case 2:
            blob = _a.sent()
            if (!blob.type.startsWith('image/')) {
              throw new Error(
                'fetchImage - expected blob type to be of type image/*, instead have: ' +
                  blob.type +
                  ', for url: ' +
                  res.url
              )
            }
            return [2 /*return*/, bufferToImage(blob)]
        }
      })
    })
  }

  function fetchJson(uri) {
    return __awaiter(this, void 0, void 0, function () {
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            return [4 /*yield*/, fetchOrThrow(uri)]
          case 1:
            return [2 /*return*/, _a.sent().json()]
        }
      })
    })
  }

  function fetchNetWeights(uri) {
    return __awaiter(this, void 0, void 0, function () {
      var _a
      return __generator(this, function (_b) {
        switch (_b.label) {
          case 0:
            _a = Float32Array.bind
            return [4 /*yield*/, fetchOrThrow(uri)]
          case 1:
            return [4 /*yield*/, _b.sent().arrayBuffer()]
          case 2:
            return [
              2 /*return*/,
              new (_a.apply(Float32Array, [void 0, _b.sent()]))(),
            ]
        }
      })
    })
  }

  function getModelUris(uri, defaultModelName) {
    var defaultManifestFilename = defaultModelName + '-weights_manifest.json'
    if (!uri) {
      return {
        modelBaseUri: '',
        manifestUri: defaultManifestFilename,
      }
    }
    if (uri === '/') {
      return {
        modelBaseUri: '/',
        manifestUri: '/' + defaultManifestFilename,
      }
    }
    var protocol = uri.startsWith('http://')
      ? 'http://'
      : uri.startsWith('https://')
      ? 'https://'
      : ''
    uri = uri.replace(protocol, '')
    var parts = uri.split('/').filter(function (s) {
      return s
    })
    var manifestFile = uri.endsWith('.json')
      ? parts[parts.length - 1]
      : defaultManifestFilename
    var modelBaseUri =
      protocol +
      (uri.endsWith('.json') ? parts.slice(0, parts.length - 1) : parts).join(
        '/'
      )
    modelBaseUri = uri.startsWith('/') ? '/' + modelBaseUri : modelBaseUri
    return {
      modelBaseUri: modelBaseUri,
      manifestUri:
        modelBaseUri === '/'
          ? '/' + manifestFile
          : modelBaseUri + '/' + manifestFile,
    }
  }

  function loadWeightMap(uri, defaultModelName) {
    return __awaiter(this, void 0, void 0, function () {
      var _a, manifestUri, modelBaseUri, manifest
      return __generator(this, function (_b) {
        switch (_b.label) {
          case 0:
            ;(_a = getModelUris(uri, defaultModelName)),
              (manifestUri = _a.manifestUri),
              (modelBaseUri = _a.modelBaseUri)
            return [4 /*yield*/, fetchJson(manifestUri)]
          case 1:
            manifest = _b.sent()
            return [2 /*return*/, ed.loadWeights(manifest, modelBaseUri)]
        }
      })
    })
  }

  function matchDimensions(input, reference, useMediaDimensions) {
    if (useMediaDimensions === void 0) {
      useMediaDimensions = false
    }
    var _a = useMediaDimensions ? getMediaDimensions(reference) : reference,
      width = _a.width,
      height = _a.height
    input.width = width
    input.height = height
    return { width: width, height: height }
  }

  var NeuralNetwork = /** @class */ (function () {
    function NeuralNetwork(_name) {
      this._name = _name
      this._params = undefined
      this._paramMappings = []
    }
    Object.defineProperty(NeuralNetwork.prototype, 'params', {
      get: function () {
        return this._params
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NeuralNetwork.prototype, 'paramMappings', {
      get: function () {
        return this._paramMappings
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(NeuralNetwork.prototype, 'isLoaded', {
      get: function () {
        return !!this.params
      },
      enumerable: true,
      configurable: true,
    })
    NeuralNetwork.prototype.getParamFromPath = function (paramPath) {
      var _a = this.traversePropertyPath(paramPath),
        obj = _a.obj,
        objProp = _a.objProp
      return obj[objProp]
    }
    NeuralNetwork.prototype.reassignParamFromPath = function (
      paramPath,
      tensor
    ) {
      var _a = this.traversePropertyPath(paramPath),
        obj = _a.obj,
        objProp = _a.objProp
      obj[objProp].dispose()
      obj[objProp] = tensor
    }
    NeuralNetwork.prototype.getParamList = function () {
      var _this = this
      return this._paramMappings.map(function (_a) {
        var paramPath = _a.paramPath
        return {
          path: paramPath,
          tensor: _this.getParamFromPath(paramPath),
        }
      })
    }
    NeuralNetwork.prototype.getTrainableParams = function () {
      return this.getParamList().filter(function (param) {
        return param.tensor instanceof St
      })
    }
    NeuralNetwork.prototype.getFrozenParams = function () {
      return this.getParamList().filter(function (param) {
        return !(param.tensor instanceof St)
      })
    }
    NeuralNetwork.prototype.variable = function () {
      var _this = this
      this.getFrozenParams().forEach(function (_a) {
        var path = _a.path,
          tensor = _a.tensor
        _this.reassignParamFromPath(path, tensor.variable())
      })
    }
    NeuralNetwork.prototype.freeze = function () {
      var _this = this
      this.getTrainableParams().forEach(function (_a) {
        var path = _a.path,
          variable = _a.tensor
        var tensor = Fn(variable.dataSync())
        variable.dispose()
        _this.reassignParamFromPath(path, tensor)
      })
    }
    NeuralNetwork.prototype.dispose = function (throwOnRedispose) {
      if (throwOnRedispose === void 0) {
        throwOnRedispose = true
      }
      this.getParamList().forEach(function (param) {
        if (throwOnRedispose && param.tensor.isDisposed) {
          throw new Error(
            'param tensor has already been disposed for path ' + param.path
          )
        }
        param.tensor.dispose()
      })
      this._params = undefined
    }
    NeuralNetwork.prototype.serializeParams = function () {
      return new Float32Array(
        this.getParamList()
          .map(function (_a) {
            var tensor = _a.tensor
            return Array.from(tensor.dataSync())
          })
          .reduce(function (flat, arr) {
            return flat.concat(arr)
          })
      )
    }
    NeuralNetwork.prototype.load = function (weightsOrUrl) {
      return __awaiter(this, void 0, void 0, function () {
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              if (weightsOrUrl instanceof Float32Array) {
                this.extractWeights(weightsOrUrl)
                return [2 /*return*/]
              }
              return [4 /*yield*/, this.loadFromUri(weightsOrUrl)]
            case 1:
              _a.sent()
              return [2 /*return*/]
          }
        })
      })
    }
    NeuralNetwork.prototype.loadFromUri = function (uri) {
      return __awaiter(this, void 0, void 0, function () {
        var weightMap
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              if (uri && typeof uri !== 'string') {
                throw new Error(
                  this._name + '.loadFromUri - expected model uri'
                )
              }
              return [
                4 /*yield*/,
                loadWeightMap(uri, this.getDefaultModelName()),
              ]
            case 1:
              weightMap = _a.sent()
              this.loadFromWeightMap(weightMap)
              return [2 /*return*/]
          }
        })
      })
    }
    NeuralNetwork.prototype.loadFromDisk = function (filePath) {
      return __awaiter(this, void 0, void 0, function () {
        var readFile,
          _a,
          manifestUri,
          modelBaseUri,
          fetchWeightsFromDisk,
          loadWeights,
          manifest,
          _b,
          _c,
          weightMap
        return __generator(this, function (_d) {
          switch (_d.label) {
            case 0:
              if (filePath && typeof filePath !== 'string') {
                throw new Error(
                  this._name + '.loadFromDisk - expected model file path'
                )
              }
              readFile = env.getEnv().readFile
              ;(_a = getModelUris(filePath, this.getDefaultModelName())),
                (manifestUri = _a.manifestUri),
                (modelBaseUri = _a.modelBaseUri)
              fetchWeightsFromDisk = function (filePaths) {
                return Promise.all(
                  filePaths.map(function (filePath) {
                    return readFile(filePath).then(function (buf) {
                      return buf.buffer
                    })
                  })
                )
              }
              loadWeights = ed.weightsLoaderFactory(fetchWeightsFromDisk)
              _c = (_b = JSON).parse
              return [4 /*yield*/, readFile(manifestUri)]
            case 1:
              manifest = _c.apply(_b, [_d.sent().toString()])
              return [4 /*yield*/, loadWeights(manifest, modelBaseUri)]
            case 2:
              weightMap = _d.sent()
              this.loadFromWeightMap(weightMap)
              return [2 /*return*/]
          }
        })
      })
    }
    NeuralNetwork.prototype.loadFromWeightMap = function (weightMap) {
      var _a = this.extractParamsFromWeigthMap(weightMap),
        paramMappings = _a.paramMappings,
        params = _a.params
      this._paramMappings = paramMappings
      this._params = params
    }
    NeuralNetwork.prototype.extractWeights = function (weights) {
      var _a = this.extractParams(weights),
        paramMappings = _a.paramMappings,
        params = _a.params
      this._paramMappings = paramMappings
      this._params = params
    }
    NeuralNetwork.prototype.traversePropertyPath = function (paramPath) {
      if (!this.params) {
        throw new Error('traversePropertyPath - model has no loaded params')
      }
      var result = paramPath.split('/').reduce(
        function (res, objProp) {
          if (!res.nextObj.hasOwnProperty(objProp)) {
            throw new Error(
              'traversePropertyPath - object does not have property ' +
                objProp +
                ', for path ' +
                paramPath
            )
          }
          return {
            obj: res.nextObj,
            objProp: objProp,
            nextObj: res.nextObj[objProp],
          }
        },
        { nextObj: this.params }
      )
      var obj = result.obj,
        objProp = result.objProp
      if (!obj || !objProp || !(obj[objProp] instanceof wt)) {
        throw new Error(
          'traversePropertyPath - parameter is not a tensor, for path ' +
            paramPath
        )
      }
      return { obj: obj, objProp: objProp }
    }
    return NeuralNetwork
  })()

  function depthwiseSeparableConv(x, params, stride) {
    return Ze(function () {
      var out = Jc(
        x,
        params.depthwise_filter,
        params.pointwise_filter,
        stride,
        'same'
      )
      out = rc(out, params.bias)
      return out
    })
  }

  function denseBlock3(x, denseBlockParams, isFirstLayer) {
    if (isFirstLayer === void 0) {
      isFirstLayer = false
    }
    return Ze(function () {
      var out1 = Bl(
        isFirstLayer
          ? rc(
              qc(x, denseBlockParams.conv0.filters, [2, 2], 'same'),
              denseBlockParams.conv0.bias
            )
          : depthwiseSeparableConv(x, denseBlockParams.conv0, [2, 2])
      )
      var out2 = depthwiseSeparableConv(out1, denseBlockParams.conv1, [1, 1])
      var in3 = Bl(rc(out1, out2))
      var out3 = depthwiseSeparableConv(in3, denseBlockParams.conv2, [1, 1])
      return Bl(rc(out1, rc(out2, out3)))
    })
  }
  function denseBlock4(x, denseBlockParams, isFirstLayer, isScaleDown) {
    if (isFirstLayer === void 0) {
      isFirstLayer = false
    }
    if (isScaleDown === void 0) {
      isScaleDown = true
    }
    return Ze(function () {
      var out1 = Bl(
        isFirstLayer
          ? rc(
              qc(
                x,
                denseBlockParams.conv0.filters,
                isScaleDown ? [2, 2] : [1, 1],
                'same'
              ),
              denseBlockParams.conv0.bias
            )
          : depthwiseSeparableConv(
              x,
              denseBlockParams.conv0,
              isScaleDown ? [2, 2] : [1, 1]
            )
      )
      var out2 = depthwiseSeparableConv(out1, denseBlockParams.conv1, [1, 1])
      var in3 = Bl(rc(out1, out2))
      var out3 = depthwiseSeparableConv(in3, denseBlockParams.conv2, [1, 1])
      var in4 = Bl(rc(out1, rc(out2, out3)))
      var out4 = depthwiseSeparableConv(in4, denseBlockParams.conv3, [1, 1])
      return Bl(rc(out1, rc(out2, rc(out3, out4))))
    })
  }

  function convLayer(x, params, padding, withRelu) {
    if (padding === void 0) {
      padding = 'same'
    }
    if (withRelu === void 0) {
      withRelu = false
    }
    return Ze(function () {
      var out = rc(qc(x, params.filters, [1, 1], padding), params.bias)
      return withRelu ? Bl(out) : out
    })
  }

  function disposeUnusedWeightTensors(weightMap, paramMappings) {
    Object.keys(weightMap).forEach(function (path) {
      if (
        !paramMappings.some(function (pm) {
          return pm.originalPath === path
        })
      ) {
        weightMap[path].dispose()
      }
    })
  }

  function extractConvParamsFactory(extractWeights, paramMappings) {
    return function (channelsIn, channelsOut, filterSize, mappedPrefix) {
      var filters = Ln(
        extractWeights(channelsIn * channelsOut * filterSize * filterSize),
        [filterSize, filterSize, channelsIn, channelsOut]
      )
      var bias = Mn(extractWeights(channelsOut))
      paramMappings.push(
        { paramPath: mappedPrefix + '/filters' },
        { paramPath: mappedPrefix + '/bias' }
      )
      return { filters: filters, bias: bias }
    }
  }

  function extractFCParamsFactory(extractWeights, paramMappings) {
    return function (channelsIn, channelsOut, mappedPrefix) {
      var fc_weights = Bn(extractWeights(channelsIn * channelsOut), [
        channelsIn,
        channelsOut,
      ])
      var fc_bias = Mn(extractWeights(channelsOut))
      paramMappings.push(
        { paramPath: mappedPrefix + '/weights' },
        { paramPath: mappedPrefix + '/bias' }
      )
      return {
        weights: fc_weights,
        bias: fc_bias,
      }
    }
  }

  var SeparableConvParams = /** @class */ (function () {
    function SeparableConvParams(depthwise_filter, pointwise_filter, bias) {
      this.depthwise_filter = depthwise_filter
      this.pointwise_filter = pointwise_filter
      this.bias = bias
    }
    return SeparableConvParams
  })()

  function extractSeparableConvParamsFactory(extractWeights, paramMappings) {
    return function (channelsIn, channelsOut, mappedPrefix) {
      var depthwise_filter = Ln(extractWeights(3 * 3 * channelsIn), [
        3,
        3,
        channelsIn,
        1,
      ])
      var pointwise_filter = Ln(extractWeights(channelsIn * channelsOut), [
        1,
        1,
        channelsIn,
        channelsOut,
      ])
      var bias = Mn(extractWeights(channelsOut))
      paramMappings.push(
        { paramPath: mappedPrefix + '/depthwise_filter' },
        { paramPath: mappedPrefix + '/pointwise_filter' },
        { paramPath: mappedPrefix + '/bias' }
      )
      return new SeparableConvParams(depthwise_filter, pointwise_filter, bias)
    }
  }
  function loadSeparableConvParamsFactory(extractWeightEntry) {
    return function (prefix) {
      var depthwise_filter = extractWeightEntry(prefix + '/depthwise_filter', 4)
      var pointwise_filter = extractWeightEntry(prefix + '/pointwise_filter', 4)
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return new SeparableConvParams(depthwise_filter, pointwise_filter, bias)
    }
  }

  function extractWeightEntryFactory(weightMap, paramMappings) {
    return function (originalPath, paramRank, mappedPath) {
      var tensor = weightMap[originalPath]
      if (!isTensor(tensor, paramRank)) {
        throw new Error(
          'expected weightMap[' +
            originalPath +
            '] to be a Tensor' +
            paramRank +
            'D, instead have ' +
            tensor
        )
      }
      paramMappings.push({
        originalPath: originalPath,
        paramPath: mappedPath || originalPath,
      })
      return tensor
    }
  }

  function extractWeightsFactory(weights) {
    var remainingWeights = weights
    function extractWeights(numWeights) {
      var ret = remainingWeights.slice(0, numWeights)
      remainingWeights = remainingWeights.slice(numWeights)
      return ret
    }
    function getRemainingWeights() {
      return remainingWeights
    }
    return {
      extractWeights: extractWeights,
      getRemainingWeights: getRemainingWeights,
    }
  }

  function extractorsFactory(extractWeights, paramMappings) {
    var extractConvParams = extractConvParamsFactory(
      extractWeights,
      paramMappings
    )
    var extractSeparableConvParams = extractSeparableConvParamsFactory(
      extractWeights,
      paramMappings
    )
    function extractDenseBlock3Params(
      channelsIn,
      channelsOut,
      mappedPrefix,
      isFirstLayer
    ) {
      if (isFirstLayer === void 0) {
        isFirstLayer = false
      }
      var conv0 = isFirstLayer
        ? extractConvParams(channelsIn, channelsOut, 3, mappedPrefix + '/conv0')
        : extractSeparableConvParams(
            channelsIn,
            channelsOut,
            mappedPrefix + '/conv0'
          )
      var conv1 = extractSeparableConvParams(
        channelsOut,
        channelsOut,
        mappedPrefix + '/conv1'
      )
      var conv2 = extractSeparableConvParams(
        channelsOut,
        channelsOut,
        mappedPrefix + '/conv2'
      )
      return { conv0: conv0, conv1: conv1, conv2: conv2 }
    }
    function extractDenseBlock4Params(
      channelsIn,
      channelsOut,
      mappedPrefix,
      isFirstLayer
    ) {
      if (isFirstLayer === void 0) {
        isFirstLayer = false
      }
      var _a = extractDenseBlock3Params(
          channelsIn,
          channelsOut,
          mappedPrefix,
          isFirstLayer
        ),
        conv0 = _a.conv0,
        conv1 = _a.conv1,
        conv2 = _a.conv2
      var conv3 = extractSeparableConvParams(
        channelsOut,
        channelsOut,
        mappedPrefix + '/conv3'
      )
      return { conv0: conv0, conv1: conv1, conv2: conv2, conv3: conv3 }
    }
    return {
      extractDenseBlock3Params: extractDenseBlock3Params,
      extractDenseBlock4Params: extractDenseBlock4Params,
    }
  }

  function extractParams(weights) {
    var paramMappings = []
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var extractDenseBlock4Params = extractorsFactory(
      extractWeights,
      paramMappings
    ).extractDenseBlock4Params
    var dense0 = extractDenseBlock4Params(3, 32, 'dense0', true)
    var dense1 = extractDenseBlock4Params(32, 64, 'dense1')
    var dense2 = extractDenseBlock4Params(64, 128, 'dense2')
    var dense3 = extractDenseBlock4Params(128, 256, 'dense3')
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      paramMappings: paramMappings,
      params: {
        dense0: dense0,
        dense1: dense1,
        dense2: dense2,
        dense3: dense3,
      },
    }
  }

  function loadConvParamsFactory(extractWeightEntry) {
    return function (prefix) {
      var filters = extractWeightEntry(prefix + '/filters', 4)
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return { filters: filters, bias: bias }
    }
  }

  function loadParamsFactory(weightMap, paramMappings) {
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    var extractConvParams = loadConvParamsFactory(extractWeightEntry)
    var extractSeparableConvParams =
      loadSeparableConvParamsFactory(extractWeightEntry)
    function extractDenseBlock3Params(prefix, isFirstLayer) {
      if (isFirstLayer === void 0) {
        isFirstLayer = false
      }
      var conv0 = isFirstLayer
        ? extractConvParams(prefix + '/conv0')
        : extractSeparableConvParams(prefix + '/conv0')
      var conv1 = extractSeparableConvParams(prefix + '/conv1')
      var conv2 = extractSeparableConvParams(prefix + '/conv2')
      return { conv0: conv0, conv1: conv1, conv2: conv2 }
    }
    function extractDenseBlock4Params(prefix, isFirstLayer) {
      if (isFirstLayer === void 0) {
        isFirstLayer = false
      }
      var conv0 = isFirstLayer
        ? extractConvParams(prefix + '/conv0')
        : extractSeparableConvParams(prefix + '/conv0')
      var conv1 = extractSeparableConvParams(prefix + '/conv1')
      var conv2 = extractSeparableConvParams(prefix + '/conv2')
      var conv3 = extractSeparableConvParams(prefix + '/conv3')
      return { conv0: conv0, conv1: conv1, conv2: conv2, conv3: conv3 }
    }
    return {
      extractDenseBlock3Params: extractDenseBlock3Params,
      extractDenseBlock4Params: extractDenseBlock4Params,
    }
  }

  function extractParamsFromWeigthMap(weightMap) {
    var paramMappings = []
    var extractDenseBlock4Params = loadParamsFactory(
      weightMap,
      paramMappings
    ).extractDenseBlock4Params
    var params = {
      dense0: extractDenseBlock4Params('dense0', true),
      dense1: extractDenseBlock4Params('dense1'),
      dense2: extractDenseBlock4Params('dense2'),
      dense3: extractDenseBlock4Params('dense3'),
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  var FaceFeatureExtractor = /** @class */ (function (_super) {
    __extends(FaceFeatureExtractor, _super)
    function FaceFeatureExtractor() {
      return _super.call(this, 'FaceFeatureExtractor') || this
    }
    FaceFeatureExtractor.prototype.forwardInput = function (input) {
      var params = this.params
      if (!params) {
        throw new Error('FaceFeatureExtractor - load model before inference')
      }
      return Ze(function () {
        var batchTensor = input.toBatchTensor(112, true)
        var meanRgb = [122.782, 117.001, 104.298]
        var normalized = normalize(batchTensor, meanRgb).div(On(255))
        var out = denseBlock4(normalized, params.dense0, true)
        out = denseBlock4(out, params.dense1)
        out = denseBlock4(out, params.dense2)
        out = denseBlock4(out, params.dense3)
        out = fl(out, [7, 7], [2, 2], 'valid')
        return out
      })
    }
    FaceFeatureExtractor.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    FaceFeatureExtractor.prototype.getDefaultModelName = function () {
      return 'face_feature_extractor_model'
    }
    FaceFeatureExtractor.prototype.extractParamsFromWeigthMap = function (
      weightMap
    ) {
      return extractParamsFromWeigthMap(weightMap)
    }
    FaceFeatureExtractor.prototype.extractParams = function (weights) {
      return extractParams(weights)
    }
    return FaceFeatureExtractor
  })(NeuralNetwork)

  function fullyConnectedLayer(x, params) {
    return Ze(function () {
      return rc(el(x, params.weights), params.bias)
    })
  }

  function extractParams$1(weights, channelsIn, channelsOut) {
    var paramMappings = []
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var extractFCParams = extractFCParamsFactory(extractWeights, paramMappings)
    var fc = extractFCParams(channelsIn, channelsOut, 'fc')
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      paramMappings: paramMappings,
      params: { fc: fc },
    }
  }

  function extractParamsFromWeigthMap$1(weightMap) {
    var paramMappings = []
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    function extractFcParams(prefix) {
      var weights = extractWeightEntry(prefix + '/weights', 2)
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return { weights: weights, bias: bias }
    }
    var params = {
      fc: extractFcParams('fc'),
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  function seperateWeightMaps(weightMap) {
    var featureExtractorMap = {}
    var classifierMap = {}
    Object.keys(weightMap).forEach(function (key) {
      var map = key.startsWith('fc') ? classifierMap : featureExtractorMap
      map[key] = weightMap[key]
    })
    return {
      featureExtractorMap: featureExtractorMap,
      classifierMap: classifierMap,
    }
  }

  var FaceProcessor = /** @class */ (function (_super) {
    __extends(FaceProcessor, _super)
    function FaceProcessor(_name, faceFeatureExtractor) {
      var _this = _super.call(this, _name) || this
      _this._faceFeatureExtractor = faceFeatureExtractor
      return _this
    }
    Object.defineProperty(FaceProcessor.prototype, 'faceFeatureExtractor', {
      get: function () {
        return this._faceFeatureExtractor
      },
      enumerable: true,
      configurable: true,
    })
    FaceProcessor.prototype.runNet = function (input) {
      var _this = this
      var params = this.params
      if (!params) {
        throw new Error(this._name + ' - load model before inference')
      }
      return Ze(function () {
        var bottleneckFeatures =
          input instanceof NetInput
            ? _this.faceFeatureExtractor.forwardInput(input)
            : input
        return fullyConnectedLayer(
          bottleneckFeatures.as2D(bottleneckFeatures.shape[0], -1),
          params.fc
        )
      })
    }
    FaceProcessor.prototype.dispose = function (throwOnRedispose) {
      if (throwOnRedispose === void 0) {
        throwOnRedispose = true
      }
      this.faceFeatureExtractor.dispose(throwOnRedispose)
      _super.prototype.dispose.call(this, throwOnRedispose)
    }
    FaceProcessor.prototype.loadClassifierParams = function (weights) {
      var _a = this.extractClassifierParams(weights),
        params = _a.params,
        paramMappings = _a.paramMappings
      this._params = params
      this._paramMappings = paramMappings
    }
    FaceProcessor.prototype.extractClassifierParams = function (weights) {
      return extractParams$1(
        weights,
        this.getClassifierChannelsIn(),
        this.getClassifierChannelsOut()
      )
    }
    FaceProcessor.prototype.extractParamsFromWeigthMap = function (weightMap) {
      var _a = seperateWeightMaps(weightMap),
        featureExtractorMap = _a.featureExtractorMap,
        classifierMap = _a.classifierMap
      this.faceFeatureExtractor.loadFromWeightMap(featureExtractorMap)
      return extractParamsFromWeigthMap$1(classifierMap)
    }
    FaceProcessor.prototype.extractParams = function (weights) {
      var cIn = this.getClassifierChannelsIn()
      var cOut = this.getClassifierChannelsOut()
      var classifierWeightSize = cOut * cIn + cOut
      var featureExtractorWeights = weights.slice(
        0,
        weights.length - classifierWeightSize
      )
      var classifierWeights = weights.slice(
        weights.length - classifierWeightSize
      )
      this.faceFeatureExtractor.extractWeights(featureExtractorWeights)
      return this.extractClassifierParams(classifierWeights)
    }
    return FaceProcessor
  })(NeuralNetwork)

  var FACE_EXPRESSION_LABELS = [
    'neutral',
    'happy',
    'sad',
    'angry',
    'fearful',
    'disgusted',
    'surprised',
  ]
  var FaceExpressions = /** @class */ (function () {
    function FaceExpressions(probabilities) {
      var _this = this
      if (probabilities.length !== 7) {
        throw new Error(
          'FaceExpressions.constructor - expected probabilities.length to be 7, have: ' +
            probabilities.length
        )
      }
      FACE_EXPRESSION_LABELS.forEach(function (expression, idx) {
        _this[expression] = probabilities[idx]
      })
    }
    FaceExpressions.prototype.asSortedArray = function () {
      var _this = this
      return FACE_EXPRESSION_LABELS.map(function (expression) {
        return { expression: expression, probability: _this[expression] }
      }).sort(function (e0, e1) {
        return e1.probability - e0.probability
      })
    }
    return FaceExpressions
  })()

  var FaceExpressionNet = /** @class */ (function (_super) {
    __extends(FaceExpressionNet, _super)
    function FaceExpressionNet(faceFeatureExtractor) {
      if (faceFeatureExtractor === void 0) {
        faceFeatureExtractor = new FaceFeatureExtractor()
      }
      return (
        _super.call(this, 'FaceExpressionNet', faceFeatureExtractor) || this
      )
    }
    FaceExpressionNet.prototype.forwardInput = function (input) {
      var _this = this
      return Ze(function () {
        return go(_this.runNet(input))
      })
    }
    FaceExpressionNet.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    FaceExpressionNet.prototype.predictExpressions = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var netInput, out, probabilitesByBatch, predictionsByBatch
        var _this = this
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              netInput = _a.sent()
              return [4 /*yield*/, this.forwardInput(netInput)]
            case 2:
              out = _a.sent()
              return [
                4 /*yield*/,
                Promise.all(
                  Ur(out).map(function (t) {
                    return __awaiter(_this, void 0, void 0, function () {
                      var data
                      return __generator(this, function (_a) {
                        switch (_a.label) {
                          case 0:
                            return [4 /*yield*/, t.data()]
                          case 1:
                            data = _a.sent()
                            t.dispose()
                            return [2 /*return*/, data]
                        }
                      })
                    })
                  })
                ),
              ]
            case 3:
              probabilitesByBatch = _a.sent()
              out.dispose()
              predictionsByBatch = probabilitesByBatch.map(function (
                probabilites
              ) {
                return new FaceExpressions(probabilites)
              })
              return [
                2 /*return*/,
                netInput.isBatchInput
                  ? predictionsByBatch
                  : predictionsByBatch[0],
              ]
          }
        })
      })
    }
    FaceExpressionNet.prototype.getDefaultModelName = function () {
      return 'face_expression_model'
    }
    FaceExpressionNet.prototype.getClassifierChannelsIn = function () {
      return 256
    }
    FaceExpressionNet.prototype.getClassifierChannelsOut = function () {
      return 7
    }
    return FaceExpressionNet
  })(FaceProcessor)

  function isWithFaceExpressions(obj) {
    return obj['expressions'] instanceof FaceExpressions
  }
  function extendWithFaceExpressions(sourceObj, expressions) {
    var extension = { expressions: expressions }
    return Object.assign({}, sourceObj, extension)
  }

  function drawFaceExpressions(
    canvasArg,
    faceExpressions,
    minConfidence,
    textFieldAnchor
  ) {
    if (minConfidence === void 0) {
      minConfidence = 0.1
    }
    var faceExpressionsArray = Array.isArray(faceExpressions)
      ? faceExpressions
      : [faceExpressions]
    faceExpressionsArray.forEach(function (e) {
      var expr =
        e instanceof FaceExpressions
          ? e
          : isWithFaceExpressions(e)
          ? e.expressions
          : undefined
      if (!expr) {
        throw new Error(
          'drawFaceExpressions - expected faceExpressions to be FaceExpressions | WithFaceExpressions<{}> or array thereof'
        )
      }
      var sorted = expr.asSortedArray()
      var resultsToDisplay = sorted.filter(function (expr) {
        return expr.probability > minConfidence
      })
      var anchor = isWithFaceDetection(e)
        ? e.detection.box.bottomLeft
        : textFieldAnchor || new Point(0, 0)
      var drawTextField = new DrawTextField(
        resultsToDisplay.map(function (expr) {
          return expr.expression + ' (' + round(expr.probability) + ')'
        }),
        anchor
      )
      drawTextField.draw(canvasArg)
    })
  }

  function isWithFaceLandmarks(obj) {
    return (
      isWithFaceDetection(obj) &&
      obj['landmarks'] instanceof FaceLandmarks &&
      obj['unshiftedLandmarks'] instanceof FaceLandmarks &&
      obj['alignedRect'] instanceof FaceDetection
    )
  }
  function extendWithFaceLandmarks(sourceObj, unshiftedLandmarks) {
    var shift = sourceObj.detection.box
    var landmarks = unshiftedLandmarks.shiftBy(shift.x, shift.y)
    var rect = landmarks.align()
    var imageDims = sourceObj.detection.imageDims
    var alignedRect = new FaceDetection(
      sourceObj.detection.score,
      rect.rescale(imageDims.reverse()),
      imageDims
    )
    var extension = {
      landmarks: landmarks,
      unshiftedLandmarks: unshiftedLandmarks,
      alignedRect: alignedRect,
    }
    return Object.assign({}, sourceObj, extension)
  }

  var DrawFaceLandmarksOptions = /** @class */ (function () {
    function DrawFaceLandmarksOptions(options) {
      if (options === void 0) {
        options = {}
      }
      var _a = options.drawLines,
        drawLines = _a === void 0 ? true : _a,
        _b = options.drawPoints,
        drawPoints = _b === void 0 ? true : _b,
        lineWidth = options.lineWidth,
        lineColor = options.lineColor,
        pointSize = options.pointSize,
        pointColor = options.pointColor
      this.drawLines = drawLines
      this.drawPoints = drawPoints
      this.lineWidth = lineWidth || 1
      this.pointSize = pointSize || 2
      this.lineColor = lineColor || 'rgba(0, 255, 255, 1)'
      this.pointColor = pointColor || 'rgba(255, 0, 255, 1)'
    }
    return DrawFaceLandmarksOptions
  })()
  var DrawFaceLandmarks = /** @class */ (function () {
    function DrawFaceLandmarks(faceLandmarks, options) {
      if (options === void 0) {
        options = {}
      }
      this.faceLandmarks = faceLandmarks
      this.options = new DrawFaceLandmarksOptions(options)
    }
    DrawFaceLandmarks.prototype.draw = function (canvasArg) {
      var ctx = getContext2dOrThrow(canvasArg)
      var _a = this.options,
        drawLines = _a.drawLines,
        drawPoints = _a.drawPoints,
        lineWidth = _a.lineWidth,
        lineColor = _a.lineColor,
        pointSize = _a.pointSize,
        pointColor = _a.pointColor
      if (drawLines && this.faceLandmarks instanceof FaceLandmarks68) {
        ctx.strokeStyle = lineColor
        ctx.lineWidth = lineWidth
        drawContour(ctx, this.faceLandmarks.getJawOutline())
        drawContour(ctx, this.faceLandmarks.getLeftEyeBrow())
        drawContour(ctx, this.faceLandmarks.getRightEyeBrow())
        drawContour(ctx, this.faceLandmarks.getNose())
        drawContour(ctx, this.faceLandmarks.getLeftEye(), true)
        drawContour(ctx, this.faceLandmarks.getRightEye(), true)
        drawContour(ctx, this.faceLandmarks.getMouth(), true)
      }
      if (drawPoints) {
        ctx.strokeStyle = pointColor
        ctx.fillStyle = pointColor
        var drawPoint = function (pt) {
          ctx.beginPath()
          ctx.arc(pt.x, pt.y, pointSize, 0, 2 * Math.PI)
          ctx.fill()
        }
        this.faceLandmarks.positions.forEach(drawPoint)
      }
    }
    return DrawFaceLandmarks
  })()
  function drawFaceLandmarks(canvasArg, faceLandmarks) {
    var faceLandmarksArray = Array.isArray(faceLandmarks)
      ? faceLandmarks
      : [faceLandmarks]
    faceLandmarksArray.forEach(function (f) {
      var landmarks =
        f instanceof FaceLandmarks
          ? f
          : isWithFaceLandmarks(f)
          ? f.landmarks
          : undefined
      if (!landmarks) {
        throw new Error(
          'drawFaceLandmarks - expected faceExpressions to be FaceLandmarks | WithFaceLandmarks<WithFaceDetection<{}>> or array thereof'
        )
      }
      new DrawFaceLandmarks(landmarks).draw(canvasArg)
    })
  }

  var index$1 = /*#__PURE__*/ Object.freeze({
    __proto__: null,
    drawContour: drawContour,
    drawDetections: drawDetections,
    drawFaceExpressions: drawFaceExpressions,
    DrawBoxOptions: DrawBoxOptions,
    DrawBox: DrawBox,
    DrawFaceLandmarksOptions: DrawFaceLandmarksOptions,
    DrawFaceLandmarks: DrawFaceLandmarks,
    drawFaceLandmarks: drawFaceLandmarks,
    get AnchorPosition() {
      return AnchorPosition
    },
    DrawTextFieldOptions: DrawTextFieldOptions,
    DrawTextField: DrawTextField,
  })

  function extractorsFactory$1(extractWeights, paramMappings) {
    var extractConvParams = extractConvParamsFactory(
      extractWeights,
      paramMappings
    )
    var extractSeparableConvParams = extractSeparableConvParamsFactory(
      extractWeights,
      paramMappings
    )
    function extractReductionBlockParams(
      channelsIn,
      channelsOut,
      mappedPrefix
    ) {
      var separable_conv0 = extractSeparableConvParams(
        channelsIn,
        channelsOut,
        mappedPrefix + '/separable_conv0'
      )
      var separable_conv1 = extractSeparableConvParams(
        channelsOut,
        channelsOut,
        mappedPrefix + '/separable_conv1'
      )
      var expansion_conv = extractConvParams(
        channelsIn,
        channelsOut,
        1,
        mappedPrefix + '/expansion_conv'
      )
      return {
        separable_conv0: separable_conv0,
        separable_conv1: separable_conv1,
        expansion_conv: expansion_conv,
      }
    }
    function extractMainBlockParams(channels, mappedPrefix) {
      var separable_conv0 = extractSeparableConvParams(
        channels,
        channels,
        mappedPrefix + '/separable_conv0'
      )
      var separable_conv1 = extractSeparableConvParams(
        channels,
        channels,
        mappedPrefix + '/separable_conv1'
      )
      var separable_conv2 = extractSeparableConvParams(
        channels,
        channels,
        mappedPrefix + '/separable_conv2'
      )
      return {
        separable_conv0: separable_conv0,
        separable_conv1: separable_conv1,
        separable_conv2: separable_conv2,
      }
    }
    return {
      extractConvParams: extractConvParams,
      extractSeparableConvParams: extractSeparableConvParams,
      extractReductionBlockParams: extractReductionBlockParams,
      extractMainBlockParams: extractMainBlockParams,
    }
  }
  function extractParams$2(weights, numMainBlocks) {
    var paramMappings = []
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var _b = extractorsFactory$1(extractWeights, paramMappings),
      extractConvParams = _b.extractConvParams,
      extractSeparableConvParams = _b.extractSeparableConvParams,
      extractReductionBlockParams = _b.extractReductionBlockParams,
      extractMainBlockParams = _b.extractMainBlockParams
    var entry_flow_conv_in = extractConvParams(3, 32, 3, 'entry_flow/conv_in')
    var entry_flow_reduction_block_0 = extractReductionBlockParams(
      32,
      64,
      'entry_flow/reduction_block_0'
    )
    var entry_flow_reduction_block_1 = extractReductionBlockParams(
      64,
      128,
      'entry_flow/reduction_block_1'
    )
    var entry_flow = {
      conv_in: entry_flow_conv_in,
      reduction_block_0: entry_flow_reduction_block_0,
      reduction_block_1: entry_flow_reduction_block_1,
    }
    var middle_flow = {}
    range(numMainBlocks, 0, 1).forEach(function (idx) {
      middle_flow['main_block_' + idx] = extractMainBlockParams(
        128,
        'middle_flow/main_block_' + idx
      )
    })
    var exit_flow_reduction_block = extractReductionBlockParams(
      128,
      256,
      'exit_flow/reduction_block'
    )
    var exit_flow_separable_conv = extractSeparableConvParams(
      256,
      512,
      'exit_flow/separable_conv'
    )
    var exit_flow = {
      reduction_block: exit_flow_reduction_block,
      separable_conv: exit_flow_separable_conv,
    }
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      paramMappings: paramMappings,
      params: {
        entry_flow: entry_flow,
        middle_flow: middle_flow,
        exit_flow: exit_flow,
      },
    }
  }

  function loadParamsFactory$1(weightMap, paramMappings) {
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    var extractConvParams = loadConvParamsFactory(extractWeightEntry)
    var extractSeparableConvParams =
      loadSeparableConvParamsFactory(extractWeightEntry)
    function extractReductionBlockParams(mappedPrefix) {
      var separable_conv0 = extractSeparableConvParams(
        mappedPrefix + '/separable_conv0'
      )
      var separable_conv1 = extractSeparableConvParams(
        mappedPrefix + '/separable_conv1'
      )
      var expansion_conv = extractConvParams(mappedPrefix + '/expansion_conv')
      return {
        separable_conv0: separable_conv0,
        separable_conv1: separable_conv1,
        expansion_conv: expansion_conv,
      }
    }
    function extractMainBlockParams(mappedPrefix) {
      var separable_conv0 = extractSeparableConvParams(
        mappedPrefix + '/separable_conv0'
      )
      var separable_conv1 = extractSeparableConvParams(
        mappedPrefix + '/separable_conv1'
      )
      var separable_conv2 = extractSeparableConvParams(
        mappedPrefix + '/separable_conv2'
      )
      return {
        separable_conv0: separable_conv0,
        separable_conv1: separable_conv1,
        separable_conv2: separable_conv2,
      }
    }
    return {
      extractConvParams: extractConvParams,
      extractSeparableConvParams: extractSeparableConvParams,
      extractReductionBlockParams: extractReductionBlockParams,
      extractMainBlockParams: extractMainBlockParams,
    }
  }
  function extractParamsFromWeigthMap$2(weightMap, numMainBlocks) {
    var paramMappings = []
    var _a = loadParamsFactory$1(weightMap, paramMappings),
      extractConvParams = _a.extractConvParams,
      extractSeparableConvParams = _a.extractSeparableConvParams,
      extractReductionBlockParams = _a.extractReductionBlockParams,
      extractMainBlockParams = _a.extractMainBlockParams
    var entry_flow_conv_in = extractConvParams('entry_flow/conv_in')
    var entry_flow_reduction_block_0 = extractReductionBlockParams(
      'entry_flow/reduction_block_0'
    )
    var entry_flow_reduction_block_1 = extractReductionBlockParams(
      'entry_flow/reduction_block_1'
    )
    var entry_flow = {
      conv_in: entry_flow_conv_in,
      reduction_block_0: entry_flow_reduction_block_0,
      reduction_block_1: entry_flow_reduction_block_1,
    }
    var middle_flow = {}
    range(numMainBlocks, 0, 1).forEach(function (idx) {
      middle_flow['main_block_' + idx] = extractMainBlockParams(
        'middle_flow/main_block_' + idx
      )
    })
    var exit_flow_reduction_block = extractReductionBlockParams(
      'exit_flow/reduction_block'
    )
    var exit_flow_separable_conv = extractSeparableConvParams(
      'exit_flow/separable_conv'
    )
    var exit_flow = {
      reduction_block: exit_flow_reduction_block,
      separable_conv: exit_flow_separable_conv,
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return {
      params: {
        entry_flow: entry_flow,
        middle_flow: middle_flow,
        exit_flow: exit_flow,
      },
      paramMappings: paramMappings,
    }
  }

  function conv(x, params, stride) {
    return rc(qc(x, params.filters, stride, 'same'), params.bias)
  }
  function reductionBlock(x, params, isActivateInput) {
    if (isActivateInput === void 0) {
      isActivateInput = true
    }
    var out = isActivateInput ? Bl(x) : x
    out = depthwiseSeparableConv(out, params.separable_conv0, [1, 1])
    out = depthwiseSeparableConv(Bl(out), params.separable_conv1, [1, 1])
    out = hl(out, [3, 3], [2, 2], 'same')
    out = rc(out, conv(x, params.expansion_conv, [2, 2]))
    return out
  }
  function mainBlock(x, params) {
    var out = depthwiseSeparableConv(Bl(x), params.separable_conv0, [1, 1])
    out = depthwiseSeparableConv(Bl(out), params.separable_conv1, [1, 1])
    out = depthwiseSeparableConv(Bl(out), params.separable_conv2, [1, 1])
    out = rc(out, x)
    return out
  }
  var TinyXception = /** @class */ (function (_super) {
    __extends(TinyXception, _super)
    function TinyXception(numMainBlocks) {
      var _this = _super.call(this, 'TinyXception') || this
      _this._numMainBlocks = numMainBlocks
      return _this
    }
    TinyXception.prototype.forwardInput = function (input) {
      var _this = this
      var params = this.params
      if (!params) {
        throw new Error('TinyXception - load model before inference')
      }
      return Ze(function () {
        var batchTensor = input.toBatchTensor(112, true)
        var meanRgb = [122.782, 117.001, 104.298]
        var normalized = normalize(batchTensor, meanRgb).div(On(256))
        var out = Bl(conv(normalized, params.entry_flow.conv_in, [2, 2]))
        out = reductionBlock(out, params.entry_flow.reduction_block_0, false)
        out = reductionBlock(out, params.entry_flow.reduction_block_1)
        range(_this._numMainBlocks, 0, 1).forEach(function (idx) {
          out = mainBlock(out, params.middle_flow['main_block_' + idx])
        })
        out = reductionBlock(out, params.exit_flow.reduction_block)
        out = Bl(
          depthwiseSeparableConv(out, params.exit_flow.separable_conv, [1, 1])
        )
        return out
      })
    }
    TinyXception.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    TinyXception.prototype.getDefaultModelName = function () {
      return 'tiny_xception_model'
    }
    TinyXception.prototype.extractParamsFromWeigthMap = function (weightMap) {
      return extractParamsFromWeigthMap$2(weightMap, this._numMainBlocks)
    }
    TinyXception.prototype.extractParams = function (weights) {
      return extractParams$2(weights, this._numMainBlocks)
    }
    return TinyXception
  })(NeuralNetwork)

  function extractParams$3(weights) {
    var paramMappings = []
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var extractFCParams = extractFCParamsFactory(extractWeights, paramMappings)
    var age = extractFCParams(512, 1, 'fc/age')
    var gender = extractFCParams(512, 2, 'fc/gender')
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      paramMappings: paramMappings,
      params: { fc: { age: age, gender: gender } },
    }
  }

  function extractParamsFromWeigthMap$3(weightMap) {
    var paramMappings = []
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    function extractFcParams(prefix) {
      var weights = extractWeightEntry(prefix + '/weights', 2)
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return { weights: weights, bias: bias }
    }
    var params = {
      fc: {
        age: extractFcParams('fc/age'),
        gender: extractFcParams('fc/gender'),
      },
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  ;(function (Gender) {
    Gender['FEMALE'] = 'female'
    Gender['MALE'] = 'male'
  })(exports.Gender || (exports.Gender = {}))

  var AgeGenderNet = /** @class */ (function (_super) {
    __extends(AgeGenderNet, _super)
    function AgeGenderNet(faceFeatureExtractor) {
      if (faceFeatureExtractor === void 0) {
        faceFeatureExtractor = new TinyXception(2)
      }
      var _this = _super.call(this, 'AgeGenderNet') || this
      _this._faceFeatureExtractor = faceFeatureExtractor
      return _this
    }
    Object.defineProperty(AgeGenderNet.prototype, 'faceFeatureExtractor', {
      get: function () {
        return this._faceFeatureExtractor
      },
      enumerable: true,
      configurable: true,
    })
    AgeGenderNet.prototype.runNet = function (input) {
      var _this = this
      var params = this.params
      if (!params) {
        throw new Error(this._name + ' - load model before inference')
      }
      return Ze(function () {
        var bottleneckFeatures =
          input instanceof NetInput
            ? _this.faceFeatureExtractor.forwardInput(input)
            : input
        var pooled = fl(bottleneckFeatures, [7, 7], [2, 2], 'valid').as2D(
          bottleneckFeatures.shape[0],
          -1
        )
        var age = fullyConnectedLayer(pooled, params.fc.age).as1D()
        var gender = fullyConnectedLayer(pooled, params.fc.gender)
        return { age: age, gender: gender }
      })
    }
    AgeGenderNet.prototype.forwardInput = function (input) {
      var _this = this
      return Ze(function () {
        var _a = _this.runNet(input),
          age = _a.age,
          gender = _a.gender
        return { age: age, gender: go(gender) }
      })
    }
    AgeGenderNet.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    AgeGenderNet.prototype.predictAgeAndGender = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var netInput,
          out,
          ages,
          genders,
          ageAndGenderTensors,
          predictionsByBatch
        var _this = this
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              netInput = _a.sent()
              return [4 /*yield*/, this.forwardInput(netInput)]
            case 2:
              out = _a.sent()
              ages = Ur(out.age)
              genders = Ur(out.gender)
              ageAndGenderTensors = ages.map(function (ageTensor, i) {
                return {
                  ageTensor: ageTensor,
                  genderTensor: genders[i],
                }
              })
              return [
                4 /*yield*/,
                Promise.all(
                  ageAndGenderTensors.map(function (_a) {
                    var ageTensor = _a.ageTensor,
                      genderTensor = _a.genderTensor
                    return __awaiter(_this, void 0, void 0, function () {
                      var age, probMale, isMale, gender, genderProbability
                      return __generator(this, function (_b) {
                        switch (_b.label) {
                          case 0:
                            return [4 /*yield*/, ageTensor.data()]
                          case 1:
                            age = _b.sent()[0]
                            return [4 /*yield*/, genderTensor.data()]
                          case 2:
                            probMale = _b.sent()[0]
                            isMale = probMale > 0.5
                            gender = isMale
                              ? exports.Gender.MALE
                              : exports.Gender.FEMALE
                            genderProbability = isMale ? probMale : 1 - probMale
                            ageTensor.dispose()
                            genderTensor.dispose()
                            return [
                              2 /*return*/,
                              {
                                age: age,
                                gender: gender,
                                genderProbability: genderProbability,
                              },
                            ]
                        }
                      })
                    })
                  })
                ),
              ]
            case 3:
              predictionsByBatch = _a.sent()
              out.age.dispose()
              out.gender.dispose()
              return [
                2 /*return*/,
                netInput.isBatchInput
                  ? predictionsByBatch
                  : predictionsByBatch[0],
              ]
          }
        })
      })
    }
    AgeGenderNet.prototype.getDefaultModelName = function () {
      return 'age_gender_model'
    }
    AgeGenderNet.prototype.dispose = function (throwOnRedispose) {
      if (throwOnRedispose === void 0) {
        throwOnRedispose = true
      }
      this.faceFeatureExtractor.dispose(throwOnRedispose)
      _super.prototype.dispose.call(this, throwOnRedispose)
    }
    AgeGenderNet.prototype.loadClassifierParams = function (weights) {
      var _a = this.extractClassifierParams(weights),
        params = _a.params,
        paramMappings = _a.paramMappings
      this._params = params
      this._paramMappings = paramMappings
    }
    AgeGenderNet.prototype.extractClassifierParams = function (weights) {
      return extractParams$3(weights)
    }
    AgeGenderNet.prototype.extractParamsFromWeigthMap = function (weightMap) {
      var _a = seperateWeightMaps(weightMap),
        featureExtractorMap = _a.featureExtractorMap,
        classifierMap = _a.classifierMap
      this.faceFeatureExtractor.loadFromWeightMap(featureExtractorMap)
      return extractParamsFromWeigthMap$3(classifierMap)
    }
    AgeGenderNet.prototype.extractParams = function (weights) {
      var classifierWeightSize = 512 * 1 + 1 + (512 * 2 + 2)
      var featureExtractorWeights = weights.slice(
        0,
        weights.length - classifierWeightSize
      )
      var classifierWeights = weights.slice(
        weights.length - classifierWeightSize
      )
      this.faceFeatureExtractor.extractWeights(featureExtractorWeights)
      return this.extractClassifierParams(classifierWeights)
    }
    return AgeGenderNet
  })(NeuralNetwork)

  var FaceLandmark68NetBase = /** @class */ (function (_super) {
    __extends(FaceLandmark68NetBase, _super)
    function FaceLandmark68NetBase() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    FaceLandmark68NetBase.prototype.postProcess = function (
      output,
      inputSize,
      originalDimensions
    ) {
      var inputDimensions = originalDimensions.map(function (_a) {
        var width = _a.width,
          height = _a.height
        var scale = inputSize / Math.max(height, width)
        return {
          width: width * scale,
          height: height * scale,
        }
      })
      var batchSize = inputDimensions.length
      return Ze(function () {
        var createInterleavedTensor = function (fillX, fillY) {
          return Pr([Hn([68], fillX), Hn([68], fillY)], 1)
            .as2D(1, 136)
            .as1D()
        }
        var getPadding = function (batchIdx, cond) {
          var _a = inputDimensions[batchIdx],
            width = _a.width,
            height = _a.height
          return cond(width, height) ? Math.abs(width - height) / 2 : 0
        }
        var getPaddingX = function (batchIdx) {
          return getPadding(batchIdx, function (w, h) {
            return w < h
          })
        }
        var getPaddingY = function (batchIdx) {
          return getPadding(batchIdx, function (w, h) {
            return h < w
          })
        }
        var landmarkTensors = output
          .mul(Hn([batchSize, 136], inputSize))
          .sub(
            Pr(
              Array.from(Array(batchSize), function (_, batchIdx) {
                return createInterleavedTensor(
                  getPaddingX(batchIdx),
                  getPaddingY(batchIdx)
                )
              })
            )
          )
          .div(
            Pr(
              Array.from(Array(batchSize), function (_, batchIdx) {
                return createInterleavedTensor(
                  inputDimensions[batchIdx].width,
                  inputDimensions[batchIdx].height
                )
              })
            )
          )
        return landmarkTensors
      })
    }
    FaceLandmark68NetBase.prototype.forwardInput = function (input) {
      var _this = this
      return Ze(function () {
        var out = _this.runNet(input)
        return _this.postProcess(
          out,
          input.inputSize,
          input.inputDimensions.map(function (_a) {
            var height = _a[0],
              width = _a[1]
            return { height: height, width: width }
          })
        )
      })
    }
    FaceLandmark68NetBase.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    FaceLandmark68NetBase.prototype.detectLandmarks = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var netInput, landmarkTensors, landmarksForBatch
        var _this = this
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              netInput = _a.sent()
              landmarkTensors = Ze(function () {
                return Ur(_this.forwardInput(netInput))
              })
              return [
                4 /*yield*/,
                Promise.all(
                  landmarkTensors.map(function (landmarkTensor, batchIdx) {
                    return __awaiter(_this, void 0, void 0, function () {
                      var landmarksArray, _a, _b, xCoords, yCoords
                      return __generator(this, function (_c) {
                        switch (_c.label) {
                          case 0:
                            _b = (_a = Array).from
                            return [4 /*yield*/, landmarkTensor.data()]
                          case 1:
                            landmarksArray = _b.apply(_a, [_c.sent()])
                            xCoords = landmarksArray.filter(function (_, i) {
                              return isEven(i)
                            })
                            yCoords = landmarksArray.filter(function (_, i) {
                              return !isEven(i)
                            })
                            return [
                              2 /*return*/,
                              new FaceLandmarks68(
                                Array(68)
                                  .fill(0)
                                  .map(function (_, i) {
                                    return new Point(xCoords[i], yCoords[i])
                                  }),
                                {
                                  height: netInput.getInputHeight(batchIdx),
                                  width: netInput.getInputWidth(batchIdx),
                                }
                              ),
                            ]
                        }
                      })
                    })
                  })
                ),
              ]
            case 2:
              landmarksForBatch = _a.sent()
              landmarkTensors.forEach(function (t) {
                return t.dispose()
              })
              return [
                2 /*return*/,
                netInput.isBatchInput
                  ? landmarksForBatch
                  : landmarksForBatch[0],
              ]
          }
        })
      })
    }
    FaceLandmark68NetBase.prototype.getClassifierChannelsOut = function () {
      return 136
    }
    return FaceLandmark68NetBase
  })(FaceProcessor)

  var FaceLandmark68Net = /** @class */ (function (_super) {
    __extends(FaceLandmark68Net, _super)
    function FaceLandmark68Net(faceFeatureExtractor) {
      if (faceFeatureExtractor === void 0) {
        faceFeatureExtractor = new FaceFeatureExtractor()
      }
      return (
        _super.call(this, 'FaceLandmark68Net', faceFeatureExtractor) || this
      )
    }
    FaceLandmark68Net.prototype.getDefaultModelName = function () {
      return 'face_landmark_68_model'
    }
    FaceLandmark68Net.prototype.getClassifierChannelsIn = function () {
      return 256
    }
    return FaceLandmark68Net
  })(FaceLandmark68NetBase)

  function extractParamsFromWeigthMapTiny(weightMap) {
    var paramMappings = []
    var extractDenseBlock3Params = loadParamsFactory(
      weightMap,
      paramMappings
    ).extractDenseBlock3Params
    var params = {
      dense0: extractDenseBlock3Params('dense0', true),
      dense1: extractDenseBlock3Params('dense1'),
      dense2: extractDenseBlock3Params('dense2'),
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  function extractParamsTiny(weights) {
    var paramMappings = []
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var extractDenseBlock3Params = extractorsFactory(
      extractWeights,
      paramMappings
    ).extractDenseBlock3Params
    var dense0 = extractDenseBlock3Params(3, 32, 'dense0', true)
    var dense1 = extractDenseBlock3Params(32, 64, 'dense1')
    var dense2 = extractDenseBlock3Params(64, 128, 'dense2')
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      paramMappings: paramMappings,
      params: { dense0: dense0, dense1: dense1, dense2: dense2 },
    }
  }

  var TinyFaceFeatureExtractor = /** @class */ (function (_super) {
    __extends(TinyFaceFeatureExtractor, _super)
    function TinyFaceFeatureExtractor() {
      return _super.call(this, 'TinyFaceFeatureExtractor') || this
    }
    TinyFaceFeatureExtractor.prototype.forwardInput = function (input) {
      var params = this.params
      if (!params) {
        throw new Error(
          'TinyFaceFeatureExtractor - load model before inference'
        )
      }
      return Ze(function () {
        var batchTensor = input.toBatchTensor(112, true)
        var meanRgb = [122.782, 117.001, 104.298]
        var normalized = normalize(batchTensor, meanRgb).div(On(255))
        var out = denseBlock3(normalized, params.dense0, true)
        out = denseBlock3(out, params.dense1)
        out = denseBlock3(out, params.dense2)
        out = fl(out, [14, 14], [2, 2], 'valid')
        return out
      })
    }
    TinyFaceFeatureExtractor.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    TinyFaceFeatureExtractor.prototype.getDefaultModelName = function () {
      return 'face_feature_extractor_tiny_model'
    }
    TinyFaceFeatureExtractor.prototype.extractParamsFromWeigthMap = function (
      weightMap
    ) {
      return extractParamsFromWeigthMapTiny(weightMap)
    }
    TinyFaceFeatureExtractor.prototype.extractParams = function (weights) {
      return extractParamsTiny(weights)
    }
    return TinyFaceFeatureExtractor
  })(NeuralNetwork)

  var FaceLandmark68TinyNet = /** @class */ (function (_super) {
    __extends(FaceLandmark68TinyNet, _super)
    function FaceLandmark68TinyNet(faceFeatureExtractor) {
      if (faceFeatureExtractor === void 0) {
        faceFeatureExtractor = new TinyFaceFeatureExtractor()
      }
      return (
        _super.call(this, 'FaceLandmark68TinyNet', faceFeatureExtractor) || this
      )
    }
    FaceLandmark68TinyNet.prototype.getDefaultModelName = function () {
      return 'face_landmark_68_tiny_model'
    }
    FaceLandmark68TinyNet.prototype.getClassifierChannelsIn = function () {
      return 128
    }
    return FaceLandmark68TinyNet
  })(FaceLandmark68NetBase)

  var FaceLandmarkNet = /** @class */ (function (_super) {
    __extends(FaceLandmarkNet, _super)
    function FaceLandmarkNet() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    return FaceLandmarkNet
  })(FaceLandmark68Net)

  function scale(x, params) {
    return rc(gc(x, params.weights), params.biases)
  }

  function convLayer$1(x, params, strides, withRelu, padding) {
    if (padding === void 0) {
      padding = 'same'
    }
    var _a = params.conv,
      filters = _a.filters,
      bias = _a.bias
    var out = qc(x, filters, strides, padding)
    out = rc(out, bias)
    out = scale(out, params.scale)
    return withRelu ? Bl(out) : out
  }
  function conv$1(x, params) {
    return convLayer$1(x, params, [1, 1], true)
  }
  function convNoRelu(x, params) {
    return convLayer$1(x, params, [1, 1], false)
  }
  function convDown(x, params) {
    return convLayer$1(x, params, [2, 2], true, 'valid')
  }

  function extractorsFactory$2(extractWeights, paramMappings) {
    function extractFilterValues(numFilterValues, numFilters, filterSize) {
      var weights = extractWeights(numFilterValues)
      var depth = weights.length / (numFilters * filterSize * filterSize)
      if (isFloat(depth)) {
        throw new Error(
          'depth has to be an integer: ' +
            depth +
            ', weights.length: ' +
            weights.length +
            ', numFilters: ' +
            numFilters +
            ', filterSize: ' +
            filterSize
        )
      }
      return Ze(function () {
        return Wl(
          Ln(weights, [numFilters, depth, filterSize, filterSize]),
          [2, 3, 1, 0]
        )
      })
    }
    function extractConvParams(
      numFilterValues,
      numFilters,
      filterSize,
      mappedPrefix
    ) {
      var filters = extractFilterValues(numFilterValues, numFilters, filterSize)
      var bias = Mn(extractWeights(numFilters))
      paramMappings.push(
        { paramPath: mappedPrefix + '/filters' },
        { paramPath: mappedPrefix + '/bias' }
      )
      return { filters: filters, bias: bias }
    }
    function extractScaleLayerParams(numWeights, mappedPrefix) {
      var weights = Mn(extractWeights(numWeights))
      var biases = Mn(extractWeights(numWeights))
      paramMappings.push(
        { paramPath: mappedPrefix + '/weights' },
        { paramPath: mappedPrefix + '/biases' }
      )
      return {
        weights: weights,
        biases: biases,
      }
    }
    function extractConvLayerParams(
      numFilterValues,
      numFilters,
      filterSize,
      mappedPrefix
    ) {
      var conv = extractConvParams(
        numFilterValues,
        numFilters,
        filterSize,
        mappedPrefix + '/conv'
      )
      var scale = extractScaleLayerParams(numFilters, mappedPrefix + '/scale')
      return { conv: conv, scale: scale }
    }
    function extractResidualLayerParams(
      numFilterValues,
      numFilters,
      filterSize,
      mappedPrefix,
      isDown
    ) {
      if (isDown === void 0) {
        isDown = false
      }
      var conv1 = extractConvLayerParams(
        (isDown ? 0.5 : 1) * numFilterValues,
        numFilters,
        filterSize,
        mappedPrefix + '/conv1'
      )
      var conv2 = extractConvLayerParams(
        numFilterValues,
        numFilters,
        filterSize,
        mappedPrefix + '/conv2'
      )
      return { conv1: conv1, conv2: conv2 }
    }
    return {
      extractConvLayerParams: extractConvLayerParams,
      extractResidualLayerParams: extractResidualLayerParams,
    }
  }
  function extractParams$4(weights) {
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var paramMappings = []
    var _b = extractorsFactory$2(extractWeights, paramMappings),
      extractConvLayerParams = _b.extractConvLayerParams,
      extractResidualLayerParams = _b.extractResidualLayerParams
    var conv32_down = extractConvLayerParams(4704, 32, 7, 'conv32_down')
    var conv32_1 = extractResidualLayerParams(9216, 32, 3, 'conv32_1')
    var conv32_2 = extractResidualLayerParams(9216, 32, 3, 'conv32_2')
    var conv32_3 = extractResidualLayerParams(9216, 32, 3, 'conv32_3')
    var conv64_down = extractResidualLayerParams(
      36864,
      64,
      3,
      'conv64_down',
      true
    )
    var conv64_1 = extractResidualLayerParams(36864, 64, 3, 'conv64_1')
    var conv64_2 = extractResidualLayerParams(36864, 64, 3, 'conv64_2')
    var conv64_3 = extractResidualLayerParams(36864, 64, 3, 'conv64_3')
    var conv128_down = extractResidualLayerParams(
      147456,
      128,
      3,
      'conv128_down',
      true
    )
    var conv128_1 = extractResidualLayerParams(147456, 128, 3, 'conv128_1')
    var conv128_2 = extractResidualLayerParams(147456, 128, 3, 'conv128_2')
    var conv256_down = extractResidualLayerParams(
      589824,
      256,
      3,
      'conv256_down',
      true
    )
    var conv256_1 = extractResidualLayerParams(589824, 256, 3, 'conv256_1')
    var conv256_2 = extractResidualLayerParams(589824, 256, 3, 'conv256_2')
    var conv256_down_out = extractResidualLayerParams(
      589824,
      256,
      3,
      'conv256_down_out'
    )
    var fc = Ze(function () {
      return Wl(Bn(extractWeights(256 * 128), [128, 256]), [1, 0])
    })
    paramMappings.push({ paramPath: 'fc' })
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    var params = {
      conv32_down: conv32_down,
      conv32_1: conv32_1,
      conv32_2: conv32_2,
      conv32_3: conv32_3,
      conv64_down: conv64_down,
      conv64_1: conv64_1,
      conv64_2: conv64_2,
      conv64_3: conv64_3,
      conv128_down: conv128_down,
      conv128_1: conv128_1,
      conv128_2: conv128_2,
      conv256_down: conv256_down,
      conv256_1: conv256_1,
      conv256_2: conv256_2,
      conv256_down_out: conv256_down_out,
      fc: fc,
    }
    return { params: params, paramMappings: paramMappings }
  }

  function extractorsFactory$3(weightMap, paramMappings) {
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    function extractScaleLayerParams(prefix) {
      var weights = extractWeightEntry(prefix + '/scale/weights', 1)
      var biases = extractWeightEntry(prefix + '/scale/biases', 1)
      return { weights: weights, biases: biases }
    }
    function extractConvLayerParams(prefix) {
      var filters = extractWeightEntry(prefix + '/conv/filters', 4)
      var bias = extractWeightEntry(prefix + '/conv/bias', 1)
      var scale = extractScaleLayerParams(prefix)
      return { conv: { filters: filters, bias: bias }, scale: scale }
    }
    function extractResidualLayerParams(prefix) {
      return {
        conv1: extractConvLayerParams(prefix + '/conv1'),
        conv2: extractConvLayerParams(prefix + '/conv2'),
      }
    }
    return {
      extractConvLayerParams: extractConvLayerParams,
      extractResidualLayerParams: extractResidualLayerParams,
    }
  }
  function extractParamsFromWeigthMap$4(weightMap) {
    var paramMappings = []
    var _a = extractorsFactory$3(weightMap, paramMappings),
      extractConvLayerParams = _a.extractConvLayerParams,
      extractResidualLayerParams = _a.extractResidualLayerParams
    var conv32_down = extractConvLayerParams('conv32_down')
    var conv32_1 = extractResidualLayerParams('conv32_1')
    var conv32_2 = extractResidualLayerParams('conv32_2')
    var conv32_3 = extractResidualLayerParams('conv32_3')
    var conv64_down = extractResidualLayerParams('conv64_down')
    var conv64_1 = extractResidualLayerParams('conv64_1')
    var conv64_2 = extractResidualLayerParams('conv64_2')
    var conv64_3 = extractResidualLayerParams('conv64_3')
    var conv128_down = extractResidualLayerParams('conv128_down')
    var conv128_1 = extractResidualLayerParams('conv128_1')
    var conv128_2 = extractResidualLayerParams('conv128_2')
    var conv256_down = extractResidualLayerParams('conv256_down')
    var conv256_1 = extractResidualLayerParams('conv256_1')
    var conv256_2 = extractResidualLayerParams('conv256_2')
    var conv256_down_out = extractResidualLayerParams('conv256_down_out')
    var fc = weightMap['fc']
    paramMappings.push({ originalPath: 'fc', paramPath: 'fc' })
    if (!isTensor2D(fc)) {
      throw new Error(
        'expected weightMap[fc] to be a Tensor2D, instead have ' + fc
      )
    }
    var params = {
      conv32_down: conv32_down,
      conv32_1: conv32_1,
      conv32_2: conv32_2,
      conv32_3: conv32_3,
      conv64_down: conv64_down,
      conv64_1: conv64_1,
      conv64_2: conv64_2,
      conv64_3: conv64_3,
      conv128_down: conv128_down,
      conv128_1: conv128_1,
      conv128_2: conv128_2,
      conv256_down: conv256_down,
      conv256_1: conv256_1,
      conv256_2: conv256_2,
      conv256_down_out: conv256_down_out,
      fc: fc,
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  function residual(x, params) {
    var out = conv$1(x, params.conv1)
    out = convNoRelu(out, params.conv2)
    out = rc(out, x)
    out = Bl(out)
    return out
  }
  function residualDown(x, params) {
    var out = convDown(x, params.conv1)
    out = convNoRelu(out, params.conv2)
    var pooled = fl(x, 2, 2, 'valid')
    var zeros = Gn(pooled.shape)
    var isPad = pooled.shape[3] !== out.shape[3]
    var isAdjustShape =
      pooled.shape[1] !== out.shape[1] || pooled.shape[2] !== out.shape[2]
    if (isAdjustShape) {
      var padShapeX = __spreadArrays(out.shape)
      padShapeX[1] = 1
      var zerosW = Gn(padShapeX)
      out = Yn([out, zerosW], 1)
      var padShapeY = __spreadArrays(out.shape)
      padShapeY[2] = 1
      var zerosH = Gn(padShapeY)
      out = Yn([out, zerosH], 2)
    }
    pooled = isPad ? Yn([pooled, zeros], 3) : pooled
    out = rc(pooled, out)
    out = Bl(out)
    return out
  }

  var FaceRecognitionNet = /** @class */ (function (_super) {
    __extends(FaceRecognitionNet, _super)
    function FaceRecognitionNet() {
      return _super.call(this, 'FaceRecognitionNet') || this
    }
    FaceRecognitionNet.prototype.forwardInput = function (input) {
      var params = this.params
      if (!params) {
        throw new Error('FaceRecognitionNet - load model before inference')
      }
      return Ze(function () {
        var batchTensor = input.toBatchTensor(150, true).toFloat()
        var meanRgb = [122.782, 117.001, 104.298]
        var normalized = normalize(batchTensor, meanRgb).div(On(256))
        var out = convDown(normalized, params.conv32_down)
        out = hl(out, 3, 2, 'valid')
        out = residual(out, params.conv32_1)
        out = residual(out, params.conv32_2)
        out = residual(out, params.conv32_3)
        out = residualDown(out, params.conv64_down)
        out = residual(out, params.conv64_1)
        out = residual(out, params.conv64_2)
        out = residual(out, params.conv64_3)
        out = residualDown(out, params.conv128_down)
        out = residual(out, params.conv128_1)
        out = residual(out, params.conv128_2)
        out = residualDown(out, params.conv256_down)
        out = residual(out, params.conv256_1)
        out = residual(out, params.conv256_2)
        out = residualDown(out, params.conv256_down_out)
        var globalAvg = out.mean([1, 2])
        var fullyConnected = el(globalAvg, params.fc)
        return fullyConnected
      })
    }
    FaceRecognitionNet.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    FaceRecognitionNet.prototype.computeFaceDescriptor = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var netInput, faceDescriptorTensors, faceDescriptorsForBatch
        var _this = this
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              netInput = _a.sent()
              faceDescriptorTensors = Ze(function () {
                return Ur(_this.forwardInput(netInput))
              })
              return [
                4 /*yield*/,
                Promise.all(
                  faceDescriptorTensors.map(function (t) {
                    return t.data()
                  })
                ),
              ]
            case 2:
              faceDescriptorsForBatch = _a.sent()
              faceDescriptorTensors.forEach(function (t) {
                return t.dispose()
              })
              return [
                2 /*return*/,
                netInput.isBatchInput
                  ? faceDescriptorsForBatch
                  : faceDescriptorsForBatch[0],
              ]
          }
        })
      })
    }
    FaceRecognitionNet.prototype.getDefaultModelName = function () {
      return 'face_recognition_model'
    }
    FaceRecognitionNet.prototype.extractParamsFromWeigthMap = function (
      weightMap
    ) {
      return extractParamsFromWeigthMap$4(weightMap)
    }
    FaceRecognitionNet.prototype.extractParams = function (weights) {
      return extractParams$4(weights)
    }
    return FaceRecognitionNet
  })(NeuralNetwork)

  function createFaceRecognitionNet(weights) {
    var net = new FaceRecognitionNet()
    net.extractWeights(weights)
    return net
  }

  function extendWithFaceDescriptor(sourceObj, descriptor) {
    var extension = { descriptor: descriptor }
    return Object.assign({}, sourceObj, extension)
  }

  function isWithAge(obj) {
    return typeof obj['age'] === 'number'
  }
  function extendWithAge(sourceObj, age) {
    var extension = { age: age }
    return Object.assign({}, sourceObj, extension)
  }

  function isWithGender(obj) {
    return (
      (obj['gender'] === exports.Gender.MALE ||
        obj['gender'] === exports.Gender.FEMALE) &&
      isValidProbablitiy(obj['genderProbability'])
    )
  }
  function extendWithGender(sourceObj, gender, genderProbability) {
    var extension = { gender: gender, genderProbability: genderProbability }
    return Object.assign({}, sourceObj, extension)
  }

  var MtcnnOptions = /** @class */ (function () {
    function MtcnnOptions(_a) {
      var _b = _a === void 0 ? {} : _a,
        minFaceSize = _b.minFaceSize,
        scaleFactor = _b.scaleFactor,
        maxNumScales = _b.maxNumScales,
        scoreThresholds = _b.scoreThresholds,
        scaleSteps = _b.scaleSteps
      this._name = 'MtcnnOptions'
      this._minFaceSize = minFaceSize || 20
      this._scaleFactor = scaleFactor || 0.709
      this._maxNumScales = maxNumScales || 10
      this._scoreThresholds = scoreThresholds || [0.6, 0.7, 0.7]
      this._scaleSteps = scaleSteps
      if (typeof this._minFaceSize !== 'number' || this._minFaceSize < 0) {
        throw new Error(
          this._name + ' - expected minFaceSize to be a number > 0'
        )
      }
      if (
        typeof this._scaleFactor !== 'number' ||
        this._scaleFactor <= 0 ||
        this._scaleFactor >= 1
      ) {
        throw new Error(
          this._name + ' - expected scaleFactor to be a number between 0 and 1'
        )
      }
      if (typeof this._maxNumScales !== 'number' || this._maxNumScales < 0) {
        throw new Error(
          this._name + ' - expected maxNumScales to be a number > 0'
        )
      }
      if (
        !Array.isArray(this._scoreThresholds) ||
        this._scoreThresholds.length !== 3 ||
        this._scoreThresholds.some(function (th) {
          return typeof th !== 'number'
        })
      ) {
        throw new Error(
          this._name +
            ' - expected scoreThresholds to be an array of numbers of length 3'
        )
      }
      if (
        this._scaleSteps &&
        (!Array.isArray(this._scaleSteps) ||
          this._scaleSteps.some(function (th) {
            return typeof th !== 'number'
          }))
      ) {
        throw new Error(
          this._name + ' - expected scaleSteps to be an array of numbers'
        )
      }
    }
    Object.defineProperty(MtcnnOptions.prototype, 'minFaceSize', {
      get: function () {
        return this._minFaceSize
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(MtcnnOptions.prototype, 'scaleFactor', {
      get: function () {
        return this._scaleFactor
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(MtcnnOptions.prototype, 'maxNumScales', {
      get: function () {
        return this._maxNumScales
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(MtcnnOptions.prototype, 'scoreThresholds', {
      get: function () {
        return this._scoreThresholds
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(MtcnnOptions.prototype, 'scaleSteps', {
      get: function () {
        return this._scaleSteps
      },
      enumerable: true,
      configurable: true,
    })
    return MtcnnOptions
  })()

  function extractorsFactory$4(extractWeights, paramMappings) {
    function extractDepthwiseConvParams(numChannels, mappedPrefix) {
      var filters = Ln(extractWeights(3 * 3 * numChannels), [
        3,
        3,
        numChannels,
        1,
      ])
      var batch_norm_scale = Mn(extractWeights(numChannels))
      var batch_norm_offset = Mn(extractWeights(numChannels))
      var batch_norm_mean = Mn(extractWeights(numChannels))
      var batch_norm_variance = Mn(extractWeights(numChannels))
      paramMappings.push(
        { paramPath: mappedPrefix + '/filters' },
        { paramPath: mappedPrefix + '/batch_norm_scale' },
        { paramPath: mappedPrefix + '/batch_norm_offset' },
        { paramPath: mappedPrefix + '/batch_norm_mean' },
        { paramPath: mappedPrefix + '/batch_norm_variance' }
      )
      return {
        filters: filters,
        batch_norm_scale: batch_norm_scale,
        batch_norm_offset: batch_norm_offset,
        batch_norm_mean: batch_norm_mean,
        batch_norm_variance: batch_norm_variance,
      }
    }
    function extractConvParams(
      channelsIn,
      channelsOut,
      filterSize,
      mappedPrefix,
      isPointwiseConv
    ) {
      var filters = Ln(
        extractWeights(channelsIn * channelsOut * filterSize * filterSize),
        [filterSize, filterSize, channelsIn, channelsOut]
      )
      var bias = Mn(extractWeights(channelsOut))
      paramMappings.push(
        { paramPath: mappedPrefix + '/filters' },
        {
          paramPath:
            mappedPrefix +
            '/' +
            (isPointwiseConv ? 'batch_norm_offset' : 'bias'),
        }
      )
      return { filters: filters, bias: bias }
    }
    function extractPointwiseConvParams(
      channelsIn,
      channelsOut,
      filterSize,
      mappedPrefix
    ) {
      var _a = extractConvParams(
          channelsIn,
          channelsOut,
          filterSize,
          mappedPrefix,
          true
        ),
        filters = _a.filters,
        bias = _a.bias
      return {
        filters: filters,
        batch_norm_offset: bias,
      }
    }
    function extractConvPairParams(channelsIn, channelsOut, mappedPrefix) {
      var depthwise_conv = extractDepthwiseConvParams(
        channelsIn,
        mappedPrefix + '/depthwise_conv'
      )
      var pointwise_conv = extractPointwiseConvParams(
        channelsIn,
        channelsOut,
        1,
        mappedPrefix + '/pointwise_conv'
      )
      return { depthwise_conv: depthwise_conv, pointwise_conv: pointwise_conv }
    }
    function extractMobilenetV1Params() {
      var conv_0 = extractPointwiseConvParams(3, 32, 3, 'mobilenetv1/conv_0')
      var conv_1 = extractConvPairParams(32, 64, 'mobilenetv1/conv_1')
      var conv_2 = extractConvPairParams(64, 128, 'mobilenetv1/conv_2')
      var conv_3 = extractConvPairParams(128, 128, 'mobilenetv1/conv_3')
      var conv_4 = extractConvPairParams(128, 256, 'mobilenetv1/conv_4')
      var conv_5 = extractConvPairParams(256, 256, 'mobilenetv1/conv_5')
      var conv_6 = extractConvPairParams(256, 512, 'mobilenetv1/conv_6')
      var conv_7 = extractConvPairParams(512, 512, 'mobilenetv1/conv_7')
      var conv_8 = extractConvPairParams(512, 512, 'mobilenetv1/conv_8')
      var conv_9 = extractConvPairParams(512, 512, 'mobilenetv1/conv_9')
      var conv_10 = extractConvPairParams(512, 512, 'mobilenetv1/conv_10')
      var conv_11 = extractConvPairParams(512, 512, 'mobilenetv1/conv_11')
      var conv_12 = extractConvPairParams(512, 1024, 'mobilenetv1/conv_12')
      var conv_13 = extractConvPairParams(1024, 1024, 'mobilenetv1/conv_13')
      return {
        conv_0: conv_0,
        conv_1: conv_1,
        conv_2: conv_2,
        conv_3: conv_3,
        conv_4: conv_4,
        conv_5: conv_5,
        conv_6: conv_6,
        conv_7: conv_7,
        conv_8: conv_8,
        conv_9: conv_9,
        conv_10: conv_10,
        conv_11: conv_11,
        conv_12: conv_12,
        conv_13: conv_13,
      }
    }
    function extractPredictionLayerParams() {
      var conv_0 = extractPointwiseConvParams(
        1024,
        256,
        1,
        'prediction_layer/conv_0'
      )
      var conv_1 = extractPointwiseConvParams(
        256,
        512,
        3,
        'prediction_layer/conv_1'
      )
      var conv_2 = extractPointwiseConvParams(
        512,
        128,
        1,
        'prediction_layer/conv_2'
      )
      var conv_3 = extractPointwiseConvParams(
        128,
        256,
        3,
        'prediction_layer/conv_3'
      )
      var conv_4 = extractPointwiseConvParams(
        256,
        128,
        1,
        'prediction_layer/conv_4'
      )
      var conv_5 = extractPointwiseConvParams(
        128,
        256,
        3,
        'prediction_layer/conv_5'
      )
      var conv_6 = extractPointwiseConvParams(
        256,
        64,
        1,
        'prediction_layer/conv_6'
      )
      var conv_7 = extractPointwiseConvParams(
        64,
        128,
        3,
        'prediction_layer/conv_7'
      )
      var box_encoding_0_predictor = extractConvParams(
        512,
        12,
        1,
        'prediction_layer/box_predictor_0/box_encoding_predictor'
      )
      var class_predictor_0 = extractConvParams(
        512,
        9,
        1,
        'prediction_layer/box_predictor_0/class_predictor'
      )
      var box_encoding_1_predictor = extractConvParams(
        1024,
        24,
        1,
        'prediction_layer/box_predictor_1/box_encoding_predictor'
      )
      var class_predictor_1 = extractConvParams(
        1024,
        18,
        1,
        'prediction_layer/box_predictor_1/class_predictor'
      )
      var box_encoding_2_predictor = extractConvParams(
        512,
        24,
        1,
        'prediction_layer/box_predictor_2/box_encoding_predictor'
      )
      var class_predictor_2 = extractConvParams(
        512,
        18,
        1,
        'prediction_layer/box_predictor_2/class_predictor'
      )
      var box_encoding_3_predictor = extractConvParams(
        256,
        24,
        1,
        'prediction_layer/box_predictor_3/box_encoding_predictor'
      )
      var class_predictor_3 = extractConvParams(
        256,
        18,
        1,
        'prediction_layer/box_predictor_3/class_predictor'
      )
      var box_encoding_4_predictor = extractConvParams(
        256,
        24,
        1,
        'prediction_layer/box_predictor_4/box_encoding_predictor'
      )
      var class_predictor_4 = extractConvParams(
        256,
        18,
        1,
        'prediction_layer/box_predictor_4/class_predictor'
      )
      var box_encoding_5_predictor = extractConvParams(
        128,
        24,
        1,
        'prediction_layer/box_predictor_5/box_encoding_predictor'
      )
      var class_predictor_5 = extractConvParams(
        128,
        18,
        1,
        'prediction_layer/box_predictor_5/class_predictor'
      )
      var box_predictor_0 = {
        box_encoding_predictor: box_encoding_0_predictor,
        class_predictor: class_predictor_0,
      }
      var box_predictor_1 = {
        box_encoding_predictor: box_encoding_1_predictor,
        class_predictor: class_predictor_1,
      }
      var box_predictor_2 = {
        box_encoding_predictor: box_encoding_2_predictor,
        class_predictor: class_predictor_2,
      }
      var box_predictor_3 = {
        box_encoding_predictor: box_encoding_3_predictor,
        class_predictor: class_predictor_3,
      }
      var box_predictor_4 = {
        box_encoding_predictor: box_encoding_4_predictor,
        class_predictor: class_predictor_4,
      }
      var box_predictor_5 = {
        box_encoding_predictor: box_encoding_5_predictor,
        class_predictor: class_predictor_5,
      }
      return {
        conv_0: conv_0,
        conv_1: conv_1,
        conv_2: conv_2,
        conv_3: conv_3,
        conv_4: conv_4,
        conv_5: conv_5,
        conv_6: conv_6,
        conv_7: conv_7,
        box_predictor_0: box_predictor_0,
        box_predictor_1: box_predictor_1,
        box_predictor_2: box_predictor_2,
        box_predictor_3: box_predictor_3,
        box_predictor_4: box_predictor_4,
        box_predictor_5: box_predictor_5,
      }
    }
    return {
      extractMobilenetV1Params: extractMobilenetV1Params,
      extractPredictionLayerParams: extractPredictionLayerParams,
    }
  }
  function extractParams$5(weights) {
    var paramMappings = []
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var _b = extractorsFactory$4(extractWeights, paramMappings),
      extractMobilenetV1Params = _b.extractMobilenetV1Params,
      extractPredictionLayerParams = _b.extractPredictionLayerParams
    var mobilenetv1 = extractMobilenetV1Params()
    var prediction_layer = extractPredictionLayerParams()
    var extra_dim = Pn(extractWeights(5118 * 4), [1, 5118, 4])
    var output_layer = {
      extra_dim: extra_dim,
    }
    paramMappings.push({ paramPath: 'output_layer/extra_dim' })
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      params: {
        mobilenetv1: mobilenetv1,
        prediction_layer: prediction_layer,
        output_layer: output_layer,
      },
      paramMappings: paramMappings,
    }
  }

  function extractorsFactory$5(weightMap, paramMappings) {
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    function extractPointwiseConvParams(prefix, idx, mappedPrefix) {
      var filters = extractWeightEntry(
        prefix + '/Conv2d_' + idx + '_pointwise/weights',
        4,
        mappedPrefix + '/filters'
      )
      var batch_norm_offset = extractWeightEntry(
        prefix + '/Conv2d_' + idx + '_pointwise/convolution_bn_offset',
        1,
        mappedPrefix + '/batch_norm_offset'
      )
      return { filters: filters, batch_norm_offset: batch_norm_offset }
    }
    function extractConvPairParams(idx) {
      var mappedPrefix = 'mobilenetv1/conv_' + idx
      var prefixDepthwiseConv = 'MobilenetV1/Conv2d_' + idx + '_depthwise'
      var mappedPrefixDepthwiseConv = mappedPrefix + '/depthwise_conv'
      var mappedPrefixPointwiseConv = mappedPrefix + '/pointwise_conv'
      var filters = extractWeightEntry(
        prefixDepthwiseConv + '/depthwise_weights',
        4,
        mappedPrefixDepthwiseConv + '/filters'
      )
      var batch_norm_scale = extractWeightEntry(
        prefixDepthwiseConv + '/BatchNorm/gamma',
        1,
        mappedPrefixDepthwiseConv + '/batch_norm_scale'
      )
      var batch_norm_offset = extractWeightEntry(
        prefixDepthwiseConv + '/BatchNorm/beta',
        1,
        mappedPrefixDepthwiseConv + '/batch_norm_offset'
      )
      var batch_norm_mean = extractWeightEntry(
        prefixDepthwiseConv + '/BatchNorm/moving_mean',
        1,
        mappedPrefixDepthwiseConv + '/batch_norm_mean'
      )
      var batch_norm_variance = extractWeightEntry(
        prefixDepthwiseConv + '/BatchNorm/moving_variance',
        1,
        mappedPrefixDepthwiseConv + '/batch_norm_variance'
      )
      return {
        depthwise_conv: {
          filters: filters,
          batch_norm_scale: batch_norm_scale,
          batch_norm_offset: batch_norm_offset,
          batch_norm_mean: batch_norm_mean,
          batch_norm_variance: batch_norm_variance,
        },
        pointwise_conv: extractPointwiseConvParams(
          'MobilenetV1',
          idx,
          mappedPrefixPointwiseConv
        ),
      }
    }
    function extractMobilenetV1Params() {
      return {
        conv_0: extractPointwiseConvParams(
          'MobilenetV1',
          0,
          'mobilenetv1/conv_0'
        ),
        conv_1: extractConvPairParams(1),
        conv_2: extractConvPairParams(2),
        conv_3: extractConvPairParams(3),
        conv_4: extractConvPairParams(4),
        conv_5: extractConvPairParams(5),
        conv_6: extractConvPairParams(6),
        conv_7: extractConvPairParams(7),
        conv_8: extractConvPairParams(8),
        conv_9: extractConvPairParams(9),
        conv_10: extractConvPairParams(10),
        conv_11: extractConvPairParams(11),
        conv_12: extractConvPairParams(12),
        conv_13: extractConvPairParams(13),
      }
    }
    function extractConvParams(prefix, mappedPrefix) {
      var filters = extractWeightEntry(
        prefix + '/weights',
        4,
        mappedPrefix + '/filters'
      )
      var bias = extractWeightEntry(
        prefix + '/biases',
        1,
        mappedPrefix + '/bias'
      )
      return { filters: filters, bias: bias }
    }
    function extractBoxPredictorParams(idx) {
      var box_encoding_predictor = extractConvParams(
        'Prediction/BoxPredictor_' + idx + '/BoxEncodingPredictor',
        'prediction_layer/box_predictor_' + idx + '/box_encoding_predictor'
      )
      var class_predictor = extractConvParams(
        'Prediction/BoxPredictor_' + idx + '/ClassPredictor',
        'prediction_layer/box_predictor_' + idx + '/class_predictor'
      )
      return {
        box_encoding_predictor: box_encoding_predictor,
        class_predictor: class_predictor,
      }
    }
    function extractPredictionLayerParams() {
      return {
        conv_0: extractPointwiseConvParams(
          'Prediction',
          0,
          'prediction_layer/conv_0'
        ),
        conv_1: extractPointwiseConvParams(
          'Prediction',
          1,
          'prediction_layer/conv_1'
        ),
        conv_2: extractPointwiseConvParams(
          'Prediction',
          2,
          'prediction_layer/conv_2'
        ),
        conv_3: extractPointwiseConvParams(
          'Prediction',
          3,
          'prediction_layer/conv_3'
        ),
        conv_4: extractPointwiseConvParams(
          'Prediction',
          4,
          'prediction_layer/conv_4'
        ),
        conv_5: extractPointwiseConvParams(
          'Prediction',
          5,
          'prediction_layer/conv_5'
        ),
        conv_6: extractPointwiseConvParams(
          'Prediction',
          6,
          'prediction_layer/conv_6'
        ),
        conv_7: extractPointwiseConvParams(
          'Prediction',
          7,
          'prediction_layer/conv_7'
        ),
        box_predictor_0: extractBoxPredictorParams(0),
        box_predictor_1: extractBoxPredictorParams(1),
        box_predictor_2: extractBoxPredictorParams(2),
        box_predictor_3: extractBoxPredictorParams(3),
        box_predictor_4: extractBoxPredictorParams(4),
        box_predictor_5: extractBoxPredictorParams(5),
      }
    }
    return {
      extractMobilenetV1Params: extractMobilenetV1Params,
      extractPredictionLayerParams: extractPredictionLayerParams,
    }
  }
  function extractParamsFromWeigthMap$5(weightMap) {
    var paramMappings = []
    var _a = extractorsFactory$5(weightMap, paramMappings),
      extractMobilenetV1Params = _a.extractMobilenetV1Params,
      extractPredictionLayerParams = _a.extractPredictionLayerParams
    var extra_dim = weightMap['Output/extra_dim']
    paramMappings.push({
      originalPath: 'Output/extra_dim',
      paramPath: 'output_layer/extra_dim',
    })
    if (!isTensor3D(extra_dim)) {
      throw new Error(
        "expected weightMap['Output/extra_dim'] to be a Tensor3D, instead have " +
          extra_dim
      )
    }
    var params = {
      mobilenetv1: extractMobilenetV1Params(),
      prediction_layer: extractPredictionLayerParams(),
      output_layer: {
        extra_dim: extra_dim,
      },
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  function pointwiseConvLayer(x, params, strides) {
    return Ze(function () {
      var out = qc(x, params.filters, strides, 'same')
      out = rc(out, params.batch_norm_offset)
      return hu(out, 0, 6)
    })
  }

  var epsilon = 0.0010000000474974513
  function depthwiseConvLayer(x, params, strides) {
    return Ze(function () {
      var out = Yc(x, params.filters, strides, 'same')
      out = ju(
        out,
        params.batch_norm_mean,
        params.batch_norm_variance,
        params.batch_norm_offset,
        params.batch_norm_scale,
        epsilon
      )
      return hu(out, 0, 6)
    })
  }
  function getStridesForLayerIdx(layerIdx) {
    return [2, 4, 6, 12].some(function (idx) {
      return idx === layerIdx
    })
      ? [2, 2]
      : [1, 1]
  }
  function mobileNetV1(x, params) {
    return Ze(function () {
      var conv11 = null
      var out = pointwiseConvLayer(x, params.conv_0, [2, 2])
      var convPairParams = [
        params.conv_1,
        params.conv_2,
        params.conv_3,
        params.conv_4,
        params.conv_5,
        params.conv_6,
        params.conv_7,
        params.conv_8,
        params.conv_9,
        params.conv_10,
        params.conv_11,
        params.conv_12,
        params.conv_13,
      ]
      convPairParams.forEach(function (param, i) {
        var layerIdx = i + 1
        var depthwiseConvStrides = getStridesForLayerIdx(layerIdx)
        out = depthwiseConvLayer(
          out,
          param.depthwise_conv,
          depthwiseConvStrides
        )
        out = pointwiseConvLayer(out, param.pointwise_conv, [1, 1])
        if (layerIdx === 11) {
          conv11 = out
        }
      })
      if (conv11 === null) {
        throw new Error('mobileNetV1 - output of conv layer 11 is null')
      }
      return {
        out: out,
        conv11: conv11,
      }
    })
  }

  function nonMaxSuppression$1(
    boxes,
    scores,
    maxOutputSize,
    iouThreshold,
    scoreThreshold
  ) {
    var numBoxes = boxes.shape[0]
    var outputSize = Math.min(maxOutputSize, numBoxes)
    var candidates = scores
      .map(function (score, boxIndex) {
        return { score: score, boxIndex: boxIndex }
      })
      .filter(function (c) {
        return c.score > scoreThreshold
      })
      .sort(function (c1, c2) {
        return c2.score - c1.score
      })
    var suppressFunc = function (x) {
      return x <= iouThreshold ? 1 : 0
    }
    var selected = []
    candidates.forEach(function (c) {
      if (selected.length >= outputSize) {
        return
      }
      var originalScore = c.score
      for (var j = selected.length - 1; j >= 0; --j) {
        var iou = IOU(boxes, c.boxIndex, selected[j])
        if (iou === 0.0) {
          continue
        }
        c.score *= suppressFunc(iou)
        if (c.score <= scoreThreshold) {
          break
        }
      }
      if (originalScore === c.score) {
        selected.push(c.boxIndex)
      }
    })
    return selected
  }
  function IOU(boxes, i, j) {
    var boxesData = boxes.arraySync()
    var yminI = Math.min(boxesData[i][0], boxesData[i][2])
    var xminI = Math.min(boxesData[i][1], boxesData[i][3])
    var ymaxI = Math.max(boxesData[i][0], boxesData[i][2])
    var xmaxI = Math.max(boxesData[i][1], boxesData[i][3])
    var yminJ = Math.min(boxesData[j][0], boxesData[j][2])
    var xminJ = Math.min(boxesData[j][1], boxesData[j][3])
    var ymaxJ = Math.max(boxesData[j][0], boxesData[j][2])
    var xmaxJ = Math.max(boxesData[j][1], boxesData[j][3])
    var areaI = (ymaxI - yminI) * (xmaxI - xminI)
    var areaJ = (ymaxJ - yminJ) * (xmaxJ - xminJ)
    if (areaI <= 0 || areaJ <= 0) {
      return 0.0
    }
    var intersectionYmin = Math.max(yminI, yminJ)
    var intersectionXmin = Math.max(xminI, xminJ)
    var intersectionYmax = Math.min(ymaxI, ymaxJ)
    var intersectionXmax = Math.min(xmaxI, xmaxJ)
    var intersectionArea =
      Math.max(intersectionYmax - intersectionYmin, 0.0) *
      Math.max(intersectionXmax - intersectionXmin, 0.0)
    return intersectionArea / (areaI + areaJ - intersectionArea)
  }

  function getCenterCoordinatesAndSizesLayer(x) {
    var vec = Ur(Wl(x, [1, 0]))
    var sizes = [Cc(vec[2], vec[0]), Cc(vec[3], vec[1])]
    var centers = [
      rc(vec[0], sc(sizes[0], On(2))),
      rc(vec[1], sc(sizes[1], On(2))),
    ]
    return {
      sizes: sizes,
      centers: centers,
    }
  }
  function decodeBoxesLayer(x0, x1) {
    var _a = getCenterCoordinatesAndSizesLayer(x0),
      sizes = _a.sizes,
      centers = _a.centers
    var vec = Ur(Wl(x1, [1, 0]))
    var div0_out = sc(gc(vu(sc(vec[2], On(5))), sizes[0]), On(2))
    var add0_out = rc(gc(sc(vec[0], On(10)), sizes[0]), centers[0])
    var div1_out = sc(gc(vu(sc(vec[3], On(5))), sizes[1]), On(2))
    var add1_out = rc(gc(sc(vec[1], On(10)), sizes[1]), centers[1])
    return Wl(
      Pr([
        Cc(add0_out, div0_out),
        Cc(add1_out, div1_out),
        rc(add0_out, div0_out),
        rc(add1_out, div1_out),
      ]),
      [1, 0]
    )
  }
  function outputLayer(boxPredictions, classPredictions, params) {
    return Ze(function () {
      var batchSize = boxPredictions.shape[0]
      var boxes = decodeBoxesLayer(
        Or(Lr(params.extra_dim, [batchSize, 1, 1]), [-1, 4]),
        Or(boxPredictions, [-1, 4])
      )
      boxes = Or(boxes, [batchSize, boxes.shape[0] / batchSize, 4])
      var scoresAndClasses = Iu(ml(classPredictions, [0, 0, 1], [-1, -1, -1]))
      var scores = ml(scoresAndClasses, [0, 0, 0], [-1, -1, 1])
      scores = Or(scores, [batchSize, scores.shape[1]])
      var boxesByBatch = Ur(boxes)
      var scoresByBatch = Ur(scores)
      return {
        boxes: boxesByBatch,
        scores: scoresByBatch,
      }
    })
  }

  function boxPredictionLayer(x, params) {
    return Ze(function () {
      var batchSize = x.shape[0]
      var boxPredictionEncoding = Or(
        convLayer(x, params.box_encoding_predictor),
        [batchSize, -1, 1, 4]
      )
      var classPrediction = Or(convLayer(x, params.class_predictor), [
        batchSize,
        -1,
        3,
      ])
      return {
        boxPredictionEncoding: boxPredictionEncoding,
        classPrediction: classPrediction,
      }
    })
  }

  function predictionLayer(x, conv11, params) {
    return Ze(function () {
      var conv0 = pointwiseConvLayer(x, params.conv_0, [1, 1])
      var conv1 = pointwiseConvLayer(conv0, params.conv_1, [2, 2])
      var conv2 = pointwiseConvLayer(conv1, params.conv_2, [1, 1])
      var conv3 = pointwiseConvLayer(conv2, params.conv_3, [2, 2])
      var conv4 = pointwiseConvLayer(conv3, params.conv_4, [1, 1])
      var conv5 = pointwiseConvLayer(conv4, params.conv_5, [2, 2])
      var conv6 = pointwiseConvLayer(conv5, params.conv_6, [1, 1])
      var conv7 = pointwiseConvLayer(conv6, params.conv_7, [2, 2])
      var boxPrediction0 = boxPredictionLayer(conv11, params.box_predictor_0)
      var boxPrediction1 = boxPredictionLayer(x, params.box_predictor_1)
      var boxPrediction2 = boxPredictionLayer(conv1, params.box_predictor_2)
      var boxPrediction3 = boxPredictionLayer(conv3, params.box_predictor_3)
      var boxPrediction4 = boxPredictionLayer(conv5, params.box_predictor_4)
      var boxPrediction5 = boxPredictionLayer(conv7, params.box_predictor_5)
      var boxPredictions = Yn(
        [
          boxPrediction0.boxPredictionEncoding,
          boxPrediction1.boxPredictionEncoding,
          boxPrediction2.boxPredictionEncoding,
          boxPrediction3.boxPredictionEncoding,
          boxPrediction4.boxPredictionEncoding,
          boxPrediction5.boxPredictionEncoding,
        ],
        1
      )
      var classPredictions = Yn(
        [
          boxPrediction0.classPrediction,
          boxPrediction1.classPrediction,
          boxPrediction2.classPrediction,
          boxPrediction3.classPrediction,
          boxPrediction4.classPrediction,
          boxPrediction5.classPrediction,
        ],
        1
      )
      return {
        boxPredictions: boxPredictions,
        classPredictions: classPredictions,
      }
    })
  }

  var SsdMobilenetv1Options = /** @class */ (function () {
    function SsdMobilenetv1Options(_a) {
      var _b = _a === void 0 ? {} : _a,
        minConfidence = _b.minConfidence,
        maxResults = _b.maxResults
      this._name = 'SsdMobilenetv1Options'
      this._minConfidence = minConfidence || 0.5
      this._maxResults = maxResults || 100
      if (
        typeof this._minConfidence !== 'number' ||
        this._minConfidence <= 0 ||
        this._minConfidence >= 1
      ) {
        throw new Error(
          this._name +
            ' - expected minConfidence to be a number between 0 and 1'
        )
      }
      if (typeof this._maxResults !== 'number') {
        throw new Error(this._name + ' - expected maxResults to be a number')
      }
    }
    Object.defineProperty(SsdMobilenetv1Options.prototype, 'minConfidence', {
      get: function () {
        return this._minConfidence
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(SsdMobilenetv1Options.prototype, 'maxResults', {
      get: function () {
        return this._maxResults
      },
      enumerable: true,
      configurable: true,
    })
    return SsdMobilenetv1Options
  })()

  var SsdMobilenetv1 = /** @class */ (function (_super) {
    __extends(SsdMobilenetv1, _super)
    function SsdMobilenetv1() {
      return _super.call(this, 'SsdMobilenetv1') || this
    }
    SsdMobilenetv1.prototype.forwardInput = function (input) {
      var params = this.params
      if (!params) {
        throw new Error('SsdMobilenetv1 - load model before inference')
      }
      return Ze(function () {
        var batchTensor = input.toBatchTensor(512, false).toFloat()
        var x = Cc(gc(batchTensor, On(0.007843137718737125)), On(1))
        var features = mobileNetV1(x, params.mobilenetv1)
        var _a = predictionLayer(
            features.out,
            features.conv11,
            params.prediction_layer
          ),
          boxPredictions = _a.boxPredictions,
          classPredictions = _a.classPredictions
        return outputLayer(
          boxPredictions,
          classPredictions,
          params.output_layer
        )
      })
    }
    SsdMobilenetv1.prototype.forward = function (input) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent()])]
          }
        })
      })
    }
    SsdMobilenetv1.prototype.locateFaces = function (input, options) {
      if (options === void 0) {
        options = {}
      }
      return __awaiter(this, void 0, void 0, function () {
        var _a,
          maxResults,
          minConfidence,
          netInput,
          _b,
          _boxes,
          _scores,
          boxes,
          scores,
          i,
          scoresData,
          _c,
          _d,
          iouThreshold,
          indices,
          reshapedDims,
          inputSize,
          padX,
          padY,
          boxesData,
          results
        return __generator(this, function (_e) {
          switch (_e.label) {
            case 0:
              ;(_a = new SsdMobilenetv1Options(options)),
                (maxResults = _a.maxResults),
                (minConfidence = _a.minConfidence)
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              netInput = _e.sent()
              ;(_b = this.forwardInput(netInput)),
                (_boxes = _b.boxes),
                (_scores = _b.scores)
              boxes = _boxes[0]
              scores = _scores[0]
              for (i = 1; i < _boxes.length; i++) {
                _boxes[i].dispose()
                _scores[i].dispose()
              }
              _d = (_c = Array).from
              return [4 /*yield*/, scores.data()]
            case 2:
              scoresData = _d.apply(_c, [_e.sent()])
              iouThreshold = 0.5
              indices = nonMaxSuppression$1(
                boxes,
                scoresData,
                maxResults,
                iouThreshold,
                minConfidence
              )
              reshapedDims = netInput.getReshapedInputDimensions(0)
              inputSize = netInput.inputSize
              padX = inputSize / reshapedDims.width
              padY = inputSize / reshapedDims.height
              boxesData = boxes.arraySync()
              results = indices.map(function (idx) {
                var _a = [
                    Math.max(0, boxesData[idx][0]),
                    Math.min(1.0, boxesData[idx][2]),
                  ].map(function (val) {
                    return val * padY
                  }),
                  top = _a[0],
                  bottom = _a[1]
                var _b = [
                    Math.max(0, boxesData[idx][1]),
                    Math.min(1.0, boxesData[idx][3]),
                  ].map(function (val) {
                    return val * padX
                  }),
                  left = _b[0],
                  right = _b[1]
                return new FaceDetection(
                  scoresData[idx],
                  new Rect(left, top, right - left, bottom - top),
                  {
                    height: netInput.getInputHeight(0),
                    width: netInput.getInputWidth(0),
                  }
                )
              })
              boxes.dispose()
              scores.dispose()
              return [2 /*return*/, results]
          }
        })
      })
    }
    SsdMobilenetv1.prototype.getDefaultModelName = function () {
      return 'ssd_mobilenetv1_model'
    }
    SsdMobilenetv1.prototype.extractParamsFromWeigthMap = function (weightMap) {
      return extractParamsFromWeigthMap$5(weightMap)
    }
    SsdMobilenetv1.prototype.extractParams = function (weights) {
      return extractParams$5(weights)
    }
    return SsdMobilenetv1
  })(NeuralNetwork)

  function createSsdMobilenetv1(weights) {
    var net = new SsdMobilenetv1()
    net.extractWeights(weights)
    return net
  }
  function createFaceDetectionNet(weights) {
    return createSsdMobilenetv1(weights)
  }
  // alias for backward compatibily
  var FaceDetectionNet = /** @class */ (function (_super) {
    __extends(FaceDetectionNet, _super)
    function FaceDetectionNet() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    return FaceDetectionNet
  })(SsdMobilenetv1)

  var IOU_THRESHOLD = 0.4
  var BOX_ANCHORS = [
    new Point(0.738768, 0.874946),
    new Point(2.42204, 2.65704),
    new Point(4.30971, 7.04493),
    new Point(10.246, 4.59428),
    new Point(12.6868, 11.8741),
  ]
  var BOX_ANCHORS_SEPARABLE = [
    new Point(1.603231, 2.094468),
    new Point(6.041143, 7.080126),
    new Point(2.882459, 3.518061),
    new Point(4.266906, 5.178857),
    new Point(9.041765, 10.66308),
  ]
  var MEAN_RGB_SEPARABLE = [117.001, 114.697, 97.404]
  var DEFAULT_MODEL_NAME = 'tiny_yolov2_model'
  var DEFAULT_MODEL_NAME_SEPARABLE_CONV = 'tiny_yolov2_separable_conv_model'

  var isNumber = function (arg) {
    return typeof arg === 'number'
  }
  function validateConfig(config) {
    if (!config) {
      throw new Error('invalid config: ' + config)
    }
    if (typeof config.withSeparableConvs !== 'boolean') {
      throw new Error(
        'config.withSeparableConvs has to be a boolean, have: ' +
          config.withSeparableConvs
      )
    }
    if (
      !isNumber(config.iouThreshold) ||
      config.iouThreshold < 0 ||
      config.iouThreshold > 1.0
    ) {
      throw new Error(
        'config.iouThreshold has to be a number between [0, 1], have: ' +
          config.iouThreshold
      )
    }
    if (
      !Array.isArray(config.classes) ||
      !config.classes.length ||
      !config.classes.every(function (c) {
        return typeof c === 'string'
      })
    ) {
      throw new Error(
        'config.classes has to be an array class names: string[], have: ' +
          JSON.stringify(config.classes)
      )
    }
    if (
      !Array.isArray(config.anchors) ||
      !config.anchors.length ||
      !config.anchors
        .map(function (a) {
          return a || {}
        })
        .every(function (a) {
          return isNumber(a.x) && isNumber(a.y)
        })
    ) {
      throw new Error(
        'config.anchors has to be an array of { x: number, y: number }, have: ' +
          JSON.stringify(config.anchors)
      )
    }
    if (
      config.meanRgb &&
      (!Array.isArray(config.meanRgb) ||
        config.meanRgb.length !== 3 ||
        !config.meanRgb.every(isNumber))
    ) {
      throw new Error(
        'config.meanRgb has to be an array of shape [number, number, number], have: ' +
          JSON.stringify(config.meanRgb)
      )
    }
  }

  function leaky(x) {
    return Ze(function () {
      var min = gc(x, On(0.10000000149011612))
      return rc(Bl(Cc(x, min)), min)
      //return tf.maximum(x, min)
    })
  }

  function convWithBatchNorm(x, params) {
    return Ze(function () {
      var out = Ir(x, [
        [0, 0],
        [1, 1],
        [1, 1],
        [0, 0],
      ])
      out = qc(out, params.conv.filters, [1, 1], 'valid')
      out = Cc(out, params.bn.sub)
      out = gc(out, params.bn.truediv)
      out = rc(out, params.conv.bias)
      return leaky(out)
    })
  }

  function depthwiseSeparableConv$1(x, params) {
    return Ze(function () {
      var out = Ir(x, [
        [0, 0],
        [1, 1],
        [1, 1],
        [0, 0],
      ])
      out = Jc(
        out,
        params.depthwise_filter,
        params.pointwise_filter,
        [1, 1],
        'valid'
      )
      out = rc(out, params.bias)
      return leaky(out)
    })
  }

  function extractorsFactory$6(extractWeights, paramMappings) {
    var extractConvParams = extractConvParamsFactory(
      extractWeights,
      paramMappings
    )
    function extractBatchNormParams(size, mappedPrefix) {
      var sub = Mn(extractWeights(size))
      var truediv = Mn(extractWeights(size))
      paramMappings.push(
        { paramPath: mappedPrefix + '/sub' },
        { paramPath: mappedPrefix + '/truediv' }
      )
      return { sub: sub, truediv: truediv }
    }
    function extractConvWithBatchNormParams(
      channelsIn,
      channelsOut,
      mappedPrefix
    ) {
      var conv = extractConvParams(
        channelsIn,
        channelsOut,
        3,
        mappedPrefix + '/conv'
      )
      var bn = extractBatchNormParams(channelsOut, mappedPrefix + '/bn')
      return { conv: conv, bn: bn }
    }
    var extractSeparableConvParams = extractSeparableConvParamsFactory(
      extractWeights,
      paramMappings
    )
    return {
      extractConvParams: extractConvParams,
      extractConvWithBatchNormParams: extractConvWithBatchNormParams,
      extractSeparableConvParams: extractSeparableConvParams,
    }
  }
  function extractParams$6(weights, config, boxEncodingSize, filterSizes) {
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var paramMappings = []
    var _b = extractorsFactory$6(extractWeights, paramMappings),
      extractConvParams = _b.extractConvParams,
      extractConvWithBatchNormParams = _b.extractConvWithBatchNormParams,
      extractSeparableConvParams = _b.extractSeparableConvParams
    var params
    if (config.withSeparableConvs) {
      var s0 = filterSizes[0],
        s1 = filterSizes[1],
        s2 = filterSizes[2],
        s3 = filterSizes[3],
        s4 = filterSizes[4],
        s5 = filterSizes[5],
        s6 = filterSizes[6],
        s7 = filterSizes[7],
        s8 = filterSizes[8]
      var conv0 = config.isFirstLayerConv2d
        ? extractConvParams(s0, s1, 3, 'conv0')
        : extractSeparableConvParams(s0, s1, 'conv0')
      var conv1 = extractSeparableConvParams(s1, s2, 'conv1')
      var conv2 = extractSeparableConvParams(s2, s3, 'conv2')
      var conv3 = extractSeparableConvParams(s3, s4, 'conv3')
      var conv4 = extractSeparableConvParams(s4, s5, 'conv4')
      var conv5 = extractSeparableConvParams(s5, s6, 'conv5')
      var conv6 = s7 ? extractSeparableConvParams(s6, s7, 'conv6') : undefined
      var conv7 = s8 ? extractSeparableConvParams(s7, s8, 'conv7') : undefined
      var conv8 = extractConvParams(
        s8 || s7 || s6,
        5 * boxEncodingSize,
        1,
        'conv8'
      )
      params = {
        conv0: conv0,
        conv1: conv1,
        conv2: conv2,
        conv3: conv3,
        conv4: conv4,
        conv5: conv5,
        conv6: conv6,
        conv7: conv7,
        conv8: conv8,
      }
    } else {
      var s0 = filterSizes[0],
        s1 = filterSizes[1],
        s2 = filterSizes[2],
        s3 = filterSizes[3],
        s4 = filterSizes[4],
        s5 = filterSizes[5],
        s6 = filterSizes[6],
        s7 = filterSizes[7],
        s8 = filterSizes[8]
      var conv0 = extractConvWithBatchNormParams(s0, s1, 'conv0')
      var conv1 = extractConvWithBatchNormParams(s1, s2, 'conv1')
      var conv2 = extractConvWithBatchNormParams(s2, s3, 'conv2')
      var conv3 = extractConvWithBatchNormParams(s3, s4, 'conv3')
      var conv4 = extractConvWithBatchNormParams(s4, s5, 'conv4')
      var conv5 = extractConvWithBatchNormParams(s5, s6, 'conv5')
      var conv6 = extractConvWithBatchNormParams(s6, s7, 'conv6')
      var conv7 = extractConvWithBatchNormParams(s7, s8, 'conv7')
      var conv8 = extractConvParams(s8, 5 * boxEncodingSize, 1, 'conv8')
      params = {
        conv0: conv0,
        conv1: conv1,
        conv2: conv2,
        conv3: conv3,
        conv4: conv4,
        conv5: conv5,
        conv6: conv6,
        conv7: conv7,
        conv8: conv8,
      }
    }
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return { params: params, paramMappings: paramMappings }
  }

  function extractorsFactory$7(weightMap, paramMappings) {
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    function extractBatchNormParams(prefix) {
      var sub = extractWeightEntry(prefix + '/sub', 1)
      var truediv = extractWeightEntry(prefix + '/truediv', 1)
      return { sub: sub, truediv: truediv }
    }
    function extractConvParams(prefix) {
      var filters = extractWeightEntry(prefix + '/filters', 4)
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return { filters: filters, bias: bias }
    }
    function extractConvWithBatchNormParams(prefix) {
      var conv = extractConvParams(prefix + '/conv')
      var bn = extractBatchNormParams(prefix + '/bn')
      return { conv: conv, bn: bn }
    }
    var extractSeparableConvParams =
      loadSeparableConvParamsFactory(extractWeightEntry)
    return {
      extractConvParams: extractConvParams,
      extractConvWithBatchNormParams: extractConvWithBatchNormParams,
      extractSeparableConvParams: extractSeparableConvParams,
    }
  }
  function extractParamsFromWeigthMap$6(weightMap, config) {
    var paramMappings = []
    var _a = extractorsFactory$7(weightMap, paramMappings),
      extractConvParams = _a.extractConvParams,
      extractConvWithBatchNormParams = _a.extractConvWithBatchNormParams,
      extractSeparableConvParams = _a.extractSeparableConvParams
    var params
    if (config.withSeparableConvs) {
      var numFilters = (config.filterSizes && config.filterSizes.length) || 9
      params = {
        conv0: config.isFirstLayerConv2d
          ? extractConvParams('conv0')
          : extractSeparableConvParams('conv0'),
        conv1: extractSeparableConvParams('conv1'),
        conv2: extractSeparableConvParams('conv2'),
        conv3: extractSeparableConvParams('conv3'),
        conv4: extractSeparableConvParams('conv4'),
        conv5: extractSeparableConvParams('conv5'),
        conv6: numFilters > 7 ? extractSeparableConvParams('conv6') : undefined,
        conv7: numFilters > 8 ? extractSeparableConvParams('conv7') : undefined,
        conv8: extractConvParams('conv8'),
      }
    } else {
      params = {
        conv0: extractConvWithBatchNormParams('conv0'),
        conv1: extractConvWithBatchNormParams('conv1'),
        conv2: extractConvWithBatchNormParams('conv2'),
        conv3: extractConvWithBatchNormParams('conv3'),
        conv4: extractConvWithBatchNormParams('conv4'),
        conv5: extractConvWithBatchNormParams('conv5'),
        conv6: extractConvWithBatchNormParams('conv6'),
        conv7: extractConvWithBatchNormParams('conv7'),
        conv8: extractConvParams('conv8'),
      }
    }
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return { params: params, paramMappings: paramMappings }
  }

  ;(function (TinyYolov2SizeType) {
    TinyYolov2SizeType[(TinyYolov2SizeType['XS'] = 224)] = 'XS'
    TinyYolov2SizeType[(TinyYolov2SizeType['SM'] = 320)] = 'SM'
    TinyYolov2SizeType[(TinyYolov2SizeType['MD'] = 416)] = 'MD'
    TinyYolov2SizeType[(TinyYolov2SizeType['LG'] = 608)] = 'LG'
  })(exports.TinyYolov2SizeType || (exports.TinyYolov2SizeType = {}))
  var TinyYolov2Options = /** @class */ (function () {
    function TinyYolov2Options(_a) {
      var _b = _a === void 0 ? {} : _a,
        inputSize = _b.inputSize,
        scoreThreshold = _b.scoreThreshold
      this._name = 'TinyYolov2Options'
      this._inputSize = inputSize || 416
      this._scoreThreshold = scoreThreshold || 0.5
      if (typeof this._inputSize !== 'number' || this._inputSize % 32 !== 0) {
        throw new Error(
          this._name + ' - expected inputSize to be a number divisible by 32'
        )
      }
      if (
        typeof this._scoreThreshold !== 'number' ||
        this._scoreThreshold <= 0 ||
        this._scoreThreshold >= 1
      ) {
        throw new Error(
          this._name +
            ' - expected scoreThreshold to be a number between 0 and 1'
        )
      }
    }
    Object.defineProperty(TinyYolov2Options.prototype, 'inputSize', {
      get: function () {
        return this._inputSize
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(TinyYolov2Options.prototype, 'scoreThreshold', {
      get: function () {
        return this._scoreThreshold
      },
      enumerable: true,
      configurable: true,
    })
    return TinyYolov2Options
  })()

  var TinyYolov2Base = /** @class */ (function (_super) {
    __extends(TinyYolov2Base, _super)
    function TinyYolov2Base(config) {
      var _this = _super.call(this, 'TinyYolov2') || this
      validateConfig(config)
      _this._config = config
      return _this
    }
    Object.defineProperty(TinyYolov2Base.prototype, 'config', {
      get: function () {
        return this._config
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(TinyYolov2Base.prototype, 'withClassScores', {
      get: function () {
        return this.config.withClassScores || this.config.classes.length > 1
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(TinyYolov2Base.prototype, 'boxEncodingSize', {
      get: function () {
        return 5 + (this.withClassScores ? this.config.classes.length : 0)
      },
      enumerable: true,
      configurable: true,
    })
    TinyYolov2Base.prototype.runTinyYolov2 = function (x, params) {
      var out = convWithBatchNorm(x, params.conv0)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = convWithBatchNorm(out, params.conv1)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = convWithBatchNorm(out, params.conv2)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = convWithBatchNorm(out, params.conv3)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = convWithBatchNorm(out, params.conv4)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = convWithBatchNorm(out, params.conv5)
      out = hl(out, [2, 2], [1, 1], 'same')
      out = convWithBatchNorm(out, params.conv6)
      out = convWithBatchNorm(out, params.conv7)
      return convLayer(out, params.conv8, 'valid', false)
    }
    TinyYolov2Base.prototype.runMobilenet = function (x, params) {
      var out = this.config.isFirstLayerConv2d
        ? leaky(convLayer(x, params.conv0, 'valid', false))
        : depthwiseSeparableConv$1(x, params.conv0)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = depthwiseSeparableConv$1(out, params.conv1)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = depthwiseSeparableConv$1(out, params.conv2)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = depthwiseSeparableConv$1(out, params.conv3)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = depthwiseSeparableConv$1(out, params.conv4)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = depthwiseSeparableConv$1(out, params.conv5)
      out = hl(out, [2, 2], [1, 1], 'same')
      out = params.conv6 ? depthwiseSeparableConv$1(out, params.conv6) : out
      out = params.conv7 ? depthwiseSeparableConv$1(out, params.conv7) : out
      return convLayer(out, params.conv8, 'valid', false)
    }
    TinyYolov2Base.prototype.forwardInput = function (input, inputSize) {
      var _this = this
      var params = this.params
      if (!params) {
        throw new Error('TinyYolov2 - load model before inference')
      }
      return Ze(function () {
        var batchTensor = input.toBatchTensor(inputSize, false).toFloat()
        batchTensor = _this.config.meanRgb
          ? normalize(batchTensor, _this.config.meanRgb)
          : batchTensor
        batchTensor = batchTensor.div(On(256))
        return _this.config.withSeparableConvs
          ? _this.runMobilenet(batchTensor, params)
          : _this.runTinyYolov2(batchTensor, params)
      })
    }
    TinyYolov2Base.prototype.forward = function (input, inputSize) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [4 /*yield*/, _a.apply(this, [_b.sent(), inputSize])]
            case 2:
              return [2 /*return*/, _b.sent()]
          }
        })
      })
    }
    TinyYolov2Base.prototype.detect = function (input, forwardParams) {
      if (forwardParams === void 0) {
        forwardParams = {}
      }
      return __awaiter(this, void 0, void 0, function () {
        var _a,
          inputSize,
          scoreThreshold,
          netInput,
          out,
          out0,
          inputDimensions,
          results,
          boxes,
          scores,
          classScores,
          classNames,
          indices,
          detections
        var _this = this
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              ;(_a = new TinyYolov2Options(forwardParams)),
                (inputSize = _a.inputSize),
                (scoreThreshold = _a.scoreThreshold)
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              netInput = _b.sent()
              return [4 /*yield*/, this.forwardInput(netInput, inputSize)]
            case 2:
              out = _b.sent()
              out0 = Ze(function () {
                return Ur(out)[0].expandDims()
              })
              inputDimensions = {
                width: netInput.getInputWidth(0),
                height: netInput.getInputHeight(0),
              }
              return [
                4 /*yield*/,
                this.extractBoxes(
                  out0,
                  netInput.getReshapedInputDimensions(0),
                  scoreThreshold
                ),
              ]
            case 3:
              results = _b.sent()
              out.dispose()
              out0.dispose()
              boxes = results.map(function (res) {
                return res.box
              })
              scores = results.map(function (res) {
                return res.score
              })
              classScores = results.map(function (res) {
                return res.classScore
              })
              classNames = results.map(function (res) {
                return _this.config.classes[res.label]
              })
              indices = nonMaxSuppression(
                boxes.map(function (box) {
                  return box.rescale(inputSize)
                }),
                scores,
                this.config.iouThreshold,
                true
              )
              detections = indices.map(function (idx) {
                return new ObjectDetection(
                  scores[idx],
                  classScores[idx],
                  classNames[idx],
                  boxes[idx],
                  inputDimensions
                )
              })
              return [2 /*return*/, detections]
          }
        })
      })
    }
    TinyYolov2Base.prototype.getDefaultModelName = function () {
      return ''
    }
    TinyYolov2Base.prototype.extractParamsFromWeigthMap = function (weightMap) {
      return extractParamsFromWeigthMap$6(weightMap, this.config)
    }
    TinyYolov2Base.prototype.extractParams = function (weights) {
      var filterSizes =
        this.config.filterSizes || TinyYolov2Base.DEFAULT_FILTER_SIZES
      var numFilters = filterSizes ? filterSizes.length : undefined
      if (numFilters !== 7 && numFilters !== 8 && numFilters !== 9) {
        throw new Error(
          'TinyYolov2 - expected 7 | 8 | 9 convolutional filters, but found ' +
            numFilters +
            ' filterSizes in config'
        )
      }
      return extractParams$6(
        weights,
        this.config,
        this.boxEncodingSize,
        filterSizes
      )
    }
    TinyYolov2Base.prototype.extractBoxes = function (
      outputTensor,
      inputBlobDimensions,
      scoreThreshold
    ) {
      return __awaiter(this, void 0, void 0, function () {
        var width,
          height,
          inputSize,
          correctionFactorX,
          correctionFactorY,
          numCells,
          numBoxes,
          _a,
          boxesTensor,
          scoresTensor,
          classScoresTensor,
          results,
          scoresData,
          boxesData,
          row,
          col,
          anchor,
          score,
          ctX,
          ctY,
          width_1,
          height_1,
          x,
          y,
          pos,
          _b,
          classScore,
          label,
          _c
        var _this = this
        return __generator(this, function (_d) {
          switch (_d.label) {
            case 0:
              ;(width = inputBlobDimensions.width),
                (height = inputBlobDimensions.height)
              inputSize = Math.max(width, height)
              correctionFactorX = inputSize / width
              correctionFactorY = inputSize / height
              numCells = outputTensor.shape[1]
              numBoxes = this.config.anchors.length
              ;(_a = Ze(function () {
                var reshaped = outputTensor.reshape([
                  numCells,
                  numCells,
                  numBoxes,
                  _this.boxEncodingSize,
                ])
                var boxes = reshaped.slice(
                  [0, 0, 0, 0],
                  [numCells, numCells, numBoxes, 4]
                )
                var scores = reshaped.slice(
                  [0, 0, 0, 4],
                  [numCells, numCells, numBoxes, 1]
                )
                var classScores = _this.withClassScores
                  ? go(
                      reshaped.slice(
                        [0, 0, 0, 5],
                        [
                          numCells,
                          numCells,
                          numBoxes,
                          _this.config.classes.length,
                        ]
                      ),
                      3
                    )
                  : On(0)
                return [boxes, scores, classScores]
              })),
                (boxesTensor = _a[0]),
                (scoresTensor = _a[1]),
                (classScoresTensor = _a[2])
              results = []
              return [4 /*yield*/, scoresTensor.array()]
            case 1:
              scoresData = _d.sent()
              return [4 /*yield*/, boxesTensor.array()]
            case 2:
              boxesData = _d.sent()
              row = 0
              _d.label = 3
            case 3:
              if (!(row < numCells)) return [3 /*break*/, 12]
              col = 0
              _d.label = 4
            case 4:
              if (!(col < numCells)) return [3 /*break*/, 11]
              anchor = 0
              _d.label = 5
            case 5:
              if (!(anchor < numBoxes)) return [3 /*break*/, 10]
              score = sigmoid(scoresData[row][col][anchor][0])
              if (!(!scoreThreshold || score > scoreThreshold))
                return [3 /*break*/, 9]
              ctX =
                ((col + sigmoid(boxesData[row][col][anchor][0])) / numCells) *
                correctionFactorX
              ctY =
                ((row + sigmoid(boxesData[row][col][anchor][1])) / numCells) *
                correctionFactorY
              width_1 =
                ((Math.exp(boxesData[row][col][anchor][2]) *
                  this.config.anchors[anchor].x) /
                  numCells) *
                correctionFactorX
              height_1 =
                ((Math.exp(boxesData[row][col][anchor][3]) *
                  this.config.anchors[anchor].y) /
                  numCells) *
                correctionFactorY
              x = ctX - width_1 / 2
              y = ctY - height_1 / 2
              pos = { row: row, col: col, anchor: anchor }
              if (!this.withClassScores) return [3 /*break*/, 7]
              return [
                4 /*yield*/,
                this.extractPredictedClass(classScoresTensor, pos),
              ]
            case 6:
              _c = _d.sent()
              return [3 /*break*/, 8]
            case 7:
              _c = { classScore: 1, label: 0 }
              _d.label = 8
            case 8:
              ;(_b = _c), (classScore = _b.classScore), (label = _b.label)
              results.push(
                __assign(
                  {
                    box: new BoundingBox(x, y, x + width_1, y + height_1),
                    score: score,
                    classScore: score * classScore,
                    label: label,
                  },
                  pos
                )
              )
              _d.label = 9
            case 9:
              anchor++
              return [3 /*break*/, 5]
            case 10:
              col++
              return [3 /*break*/, 4]
            case 11:
              row++
              return [3 /*break*/, 3]
            case 12:
              boxesTensor.dispose()
              scoresTensor.dispose()
              classScoresTensor.dispose()
              return [2 /*return*/, results]
          }
        })
      })
    }
    TinyYolov2Base.prototype.extractPredictedClass = function (
      classesTensor,
      pos
    ) {
      return __awaiter(this, void 0, void 0, function () {
        var row, col, anchor, classesData
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              ;(row = pos.row), (col = pos.col), (anchor = pos.anchor)
              return [4 /*yield*/, classesTensor.array()]
            case 1:
              classesData = _a.sent()
              return [
                2 /*return*/,
                Array(this.config.classes.length)
                  .fill(0)
                  .map(function (_, i) {
                    return classesData[row][col][anchor][i]
                  })
                  .map(function (classScore, label) {
                    return {
                      classScore: classScore,
                      label: label,
                    }
                  })
                  .reduce(function (max, curr) {
                    return max.classScore > curr.classScore ? max : curr
                  }),
              ]
          }
        })
      })
    }
    TinyYolov2Base.DEFAULT_FILTER_SIZES = [
      3, 16, 32, 64, 128, 256, 512, 1024, 1024,
    ]
    return TinyYolov2Base
  })(NeuralNetwork)

  var TinyYolov2 = /** @class */ (function (_super) {
    __extends(TinyYolov2, _super)
    function TinyYolov2(withSeparableConvs) {
      if (withSeparableConvs === void 0) {
        withSeparableConvs = true
      }
      var _this = this
      var config = Object.assign(
        {},
        {
          withSeparableConvs: withSeparableConvs,
          iouThreshold: IOU_THRESHOLD,
          classes: ['face'],
        },
        withSeparableConvs
          ? {
              anchors: BOX_ANCHORS_SEPARABLE,
              meanRgb: MEAN_RGB_SEPARABLE,
            }
          : {
              anchors: BOX_ANCHORS,
              withClassScores: true,
            }
      )
      _this = _super.call(this, config) || this
      return _this
    }
    Object.defineProperty(TinyYolov2.prototype, 'withSeparableConvs', {
      get: function () {
        return this.config.withSeparableConvs
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(TinyYolov2.prototype, 'anchors', {
      get: function () {
        return this.config.anchors
      },
      enumerable: true,
      configurable: true,
    })
    TinyYolov2.prototype.locateFaces = function (input, forwardParams) {
      return __awaiter(this, void 0, void 0, function () {
        var objectDetections
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.detect(input, forwardParams)]
            case 1:
              objectDetections = _a.sent()
              return [
                2 /*return*/,
                objectDetections.map(function (det) {
                  return new FaceDetection(det.score, det.relativeBox, {
                    width: det.imageWidth,
                    height: det.imageHeight,
                  })
                }),
              ]
          }
        })
      })
    }
    TinyYolov2.prototype.getDefaultModelName = function () {
      return this.withSeparableConvs
        ? DEFAULT_MODEL_NAME_SEPARABLE_CONV
        : DEFAULT_MODEL_NAME
    }
    TinyYolov2.prototype.extractParamsFromWeigthMap = function (weightMap) {
      return _super.prototype.extractParamsFromWeigthMap.call(this, weightMap)
    }
    return TinyYolov2
  })(TinyYolov2Base)

  function createTinyYolov2(weights, withSeparableConvs) {
    if (withSeparableConvs === void 0) {
      withSeparableConvs = true
    }
    var net = new TinyYolov2(withSeparableConvs)
    net.extractWeights(weights)
    return net
  }

  var TinyFaceDetectorOptions = /** @class */ (function (_super) {
    __extends(TinyFaceDetectorOptions, _super)
    function TinyFaceDetectorOptions() {
      var _this = (_super !== null && _super.apply(this, arguments)) || this
      _this._name = 'TinyFaceDetectorOptions'
      return _this
    }
    return TinyFaceDetectorOptions
  })(TinyYolov2Options)

  var ComposableTask = /** @class */ (function () {
    function ComposableTask() {}
    ComposableTask.prototype.then = function (onfulfilled) {
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = onfulfilled
              return [4 /*yield*/, this.run()]
            case 1:
              return [2 /*return*/, _a.apply(void 0, [_b.sent()])]
          }
        })
      })
    }
    ComposableTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        return __generator(this, function (_a) {
          throw new Error('ComposableTask - run is not implemented')
        })
      })
    }
    return ComposableTask
  })()

  function extractAllFacesAndComputeResults(
    parentResults,
    input,
    computeResults,
    extractedFaces,
    getRectForAlignment
  ) {
    if (getRectForAlignment === void 0) {
      getRectForAlignment = function (_a) {
        var alignedRect = _a.alignedRect
        return alignedRect
      }
    }
    return __awaiter(this, void 0, void 0, function () {
      var faceBoxes, faces, _a, _b, results
      return __generator(this, function (_c) {
        switch (_c.label) {
          case 0:
            faceBoxes = parentResults.map(function (parentResult) {
              return isWithFaceLandmarks(parentResult)
                ? getRectForAlignment(parentResult)
                : parentResult.detection
            })
            _a = extractedFaces
            if (_a) return [3 /*break*/, 5]
            if (!(input instanceof wt)) return [3 /*break*/, 2]
            return [4 /*yield*/, extractFaceTensors(input, faceBoxes)]
          case 1:
            _b = _c.sent()
            return [3 /*break*/, 4]
          case 2:
            return [4 /*yield*/, extractFaces(input, faceBoxes)]
          case 3:
            _b = _c.sent()
            _c.label = 4
          case 4:
            _a = _b
            _c.label = 5
          case 5:
            faces = _a
            return [4 /*yield*/, computeResults(faces)]
          case 6:
            results = _c.sent()
            faces.forEach(function (f) {
              return f instanceof wt && f.dispose()
            })
            return [2 /*return*/, results]
        }
      })
    })
  }
  function extractSingleFaceAndComputeResult(
    parentResult,
    input,
    computeResult,
    extractedFaces,
    getRectForAlignment
  ) {
    return __awaiter(this, void 0, void 0, function () {
      var _this = this
      return __generator(this, function (_a) {
        return [
          2 /*return*/,
          extractAllFacesAndComputeResults(
            [parentResult],
            input,
            function (faces) {
              return __awaiter(_this, void 0, void 0, function () {
                return __generator(this, function (_a) {
                  return [2 /*return*/, computeResult(faces[0])]
                })
              })
            },
            extractedFaces,
            getRectForAlignment
          ),
        ]
      })
    })
  }

  function bgrToRgbTensor(tensor) {
    return Ze(function () {
      return Pr(Ur(tensor, 3).reverse(), 3)
    })
  }

  var CELL_STRIDE = 2
  var CELL_SIZE = 12

  function extractorsFactory$8(extractWeights, paramMappings) {
    var extractConvParams = extractConvParamsFactory(
      extractWeights,
      paramMappings
    )
    var extractFCParams = extractFCParamsFactory(extractWeights, paramMappings)
    function extractPReluParams(size, paramPath) {
      var alpha = Mn(extractWeights(size))
      paramMappings.push({ paramPath: paramPath })
      return alpha
    }
    function extractSharedParams(numFilters, mappedPrefix, isRnet) {
      if (isRnet === void 0) {
        isRnet = false
      }
      var conv1 = extractConvParams(
        numFilters[0],
        numFilters[1],
        3,
        mappedPrefix + '/conv1'
      )
      var prelu1_alpha = extractPReluParams(
        numFilters[1],
        mappedPrefix + '/prelu1_alpha'
      )
      var conv2 = extractConvParams(
        numFilters[1],
        numFilters[2],
        3,
        mappedPrefix + '/conv2'
      )
      var prelu2_alpha = extractPReluParams(
        numFilters[2],
        mappedPrefix + '/prelu2_alpha'
      )
      var conv3 = extractConvParams(
        numFilters[2],
        numFilters[3],
        isRnet ? 2 : 3,
        mappedPrefix + '/conv3'
      )
      var prelu3_alpha = extractPReluParams(
        numFilters[3],
        mappedPrefix + '/prelu3_alpha'
      )
      return {
        conv1: conv1,
        prelu1_alpha: prelu1_alpha,
        conv2: conv2,
        prelu2_alpha: prelu2_alpha,
        conv3: conv3,
        prelu3_alpha: prelu3_alpha,
      }
    }
    function extractPNetParams() {
      var sharedParams = extractSharedParams([3, 10, 16, 32], 'pnet')
      var conv4_1 = extractConvParams(32, 2, 1, 'pnet/conv4_1')
      var conv4_2 = extractConvParams(32, 4, 1, 'pnet/conv4_2')
      return __assign(__assign({}, sharedParams), {
        conv4_1: conv4_1,
        conv4_2: conv4_2,
      })
    }
    function extractRNetParams() {
      var sharedParams = extractSharedParams([3, 28, 48, 64], 'rnet', true)
      var fc1 = extractFCParams(576, 128, 'rnet/fc1')
      var prelu4_alpha = extractPReluParams(128, 'rnet/prelu4_alpha')
      var fc2_1 = extractFCParams(128, 2, 'rnet/fc2_1')
      var fc2_2 = extractFCParams(128, 4, 'rnet/fc2_2')
      return __assign(__assign({}, sharedParams), {
        fc1: fc1,
        prelu4_alpha: prelu4_alpha,
        fc2_1: fc2_1,
        fc2_2: fc2_2,
      })
    }
    function extractONetParams() {
      var sharedParams = extractSharedParams([3, 32, 64, 64], 'onet')
      var conv4 = extractConvParams(64, 128, 2, 'onet/conv4')
      var prelu4_alpha = extractPReluParams(128, 'onet/prelu4_alpha')
      var fc1 = extractFCParams(1152, 256, 'onet/fc1')
      var prelu5_alpha = extractPReluParams(256, 'onet/prelu5_alpha')
      var fc2_1 = extractFCParams(256, 2, 'onet/fc2_1')
      var fc2_2 = extractFCParams(256, 4, 'onet/fc2_2')
      var fc2_3 = extractFCParams(256, 10, 'onet/fc2_3')
      return __assign(__assign({}, sharedParams), {
        conv4: conv4,
        prelu4_alpha: prelu4_alpha,
        fc1: fc1,
        prelu5_alpha: prelu5_alpha,
        fc2_1: fc2_1,
        fc2_2: fc2_2,
        fc2_3: fc2_3,
      })
    }
    return {
      extractPNetParams: extractPNetParams,
      extractRNetParams: extractRNetParams,
      extractONetParams: extractONetParams,
    }
  }
  function extractParams$7(weights) {
    var _a = extractWeightsFactory(weights),
      extractWeights = _a.extractWeights,
      getRemainingWeights = _a.getRemainingWeights
    var paramMappings = []
    var _b = extractorsFactory$8(extractWeights, paramMappings),
      extractPNetParams = _b.extractPNetParams,
      extractRNetParams = _b.extractRNetParams,
      extractONetParams = _b.extractONetParams
    var pnet = extractPNetParams()
    var rnet = extractRNetParams()
    var onet = extractONetParams()
    if (getRemainingWeights().length !== 0) {
      throw new Error(
        'weights remaing after extract: ' + getRemainingWeights().length
      )
    }
    return {
      params: { pnet: pnet, rnet: rnet, onet: onet },
      paramMappings: paramMappings,
    }
  }

  function extractorsFactory$9(weightMap, paramMappings) {
    var extractWeightEntry = extractWeightEntryFactory(weightMap, paramMappings)
    function extractConvParams(prefix) {
      var filters = extractWeightEntry(
        prefix + '/weights',
        4,
        prefix + '/filters'
      )
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return { filters: filters, bias: bias }
    }
    function extractFCParams(prefix) {
      var weights = extractWeightEntry(prefix + '/weights', 2)
      var bias = extractWeightEntry(prefix + '/bias', 1)
      return { weights: weights, bias: bias }
    }
    function extractPReluParams(paramPath) {
      return extractWeightEntry(paramPath, 1)
    }
    function extractSharedParams(prefix) {
      var conv1 = extractConvParams(prefix + '/conv1')
      var prelu1_alpha = extractPReluParams(prefix + '/prelu1_alpha')
      var conv2 = extractConvParams(prefix + '/conv2')
      var prelu2_alpha = extractPReluParams(prefix + '/prelu2_alpha')
      var conv3 = extractConvParams(prefix + '/conv3')
      var prelu3_alpha = extractPReluParams(prefix + '/prelu3_alpha')
      return {
        conv1: conv1,
        prelu1_alpha: prelu1_alpha,
        conv2: conv2,
        prelu2_alpha: prelu2_alpha,
        conv3: conv3,
        prelu3_alpha: prelu3_alpha,
      }
    }
    function extractPNetParams() {
      var sharedParams = extractSharedParams('pnet')
      var conv4_1 = extractConvParams('pnet/conv4_1')
      var conv4_2 = extractConvParams('pnet/conv4_2')
      return __assign(__assign({}, sharedParams), {
        conv4_1: conv4_1,
        conv4_2: conv4_2,
      })
    }
    function extractRNetParams() {
      var sharedParams = extractSharedParams('rnet')
      var fc1 = extractFCParams('rnet/fc1')
      var prelu4_alpha = extractPReluParams('rnet/prelu4_alpha')
      var fc2_1 = extractFCParams('rnet/fc2_1')
      var fc2_2 = extractFCParams('rnet/fc2_2')
      return __assign(__assign({}, sharedParams), {
        fc1: fc1,
        prelu4_alpha: prelu4_alpha,
        fc2_1: fc2_1,
        fc2_2: fc2_2,
      })
    }
    function extractONetParams() {
      var sharedParams = extractSharedParams('onet')
      var conv4 = extractConvParams('onet/conv4')
      var prelu4_alpha = extractPReluParams('onet/prelu4_alpha')
      var fc1 = extractFCParams('onet/fc1')
      var prelu5_alpha = extractPReluParams('onet/prelu5_alpha')
      var fc2_1 = extractFCParams('onet/fc2_1')
      var fc2_2 = extractFCParams('onet/fc2_2')
      var fc2_3 = extractFCParams('onet/fc2_3')
      return __assign(__assign({}, sharedParams), {
        conv4: conv4,
        prelu4_alpha: prelu4_alpha,
        fc1: fc1,
        prelu5_alpha: prelu5_alpha,
        fc2_1: fc2_1,
        fc2_2: fc2_2,
        fc2_3: fc2_3,
      })
    }
    return {
      extractPNetParams: extractPNetParams,
      extractRNetParams: extractRNetParams,
      extractONetParams: extractONetParams,
    }
  }
  function extractParamsFromWeigthMap$7(weightMap) {
    var paramMappings = []
    var _a = extractorsFactory$9(weightMap, paramMappings),
      extractPNetParams = _a.extractPNetParams,
      extractRNetParams = _a.extractRNetParams,
      extractONetParams = _a.extractONetParams
    var pnet = extractPNetParams()
    var rnet = extractRNetParams()
    var onet = extractONetParams()
    disposeUnusedWeightTensors(weightMap, paramMappings)
    return {
      params: { pnet: pnet, rnet: rnet, onet: onet },
      paramMappings: paramMappings,
    }
  }

  function getSizesForScale(scale, _a) {
    var height = _a[0],
      width = _a[1]
    return {
      height: Math.floor(height * scale),
      width: Math.floor(width * scale),
    }
  }

  function pyramidDown(minFaceSize, scaleFactor, dims) {
    var height = dims[0],
      width = dims[1]
    var m = CELL_SIZE / minFaceSize
    var scales = []
    var minLayer = Math.min(height, width) * m
    var exp = 0
    while (minLayer >= 12) {
      scales.push(m * Math.pow(scaleFactor, exp))
      minLayer = minLayer * scaleFactor
      exp += 1
    }
    return scales
  }

  var MtcnnBox = /** @class */ (function (_super) {
    __extends(MtcnnBox, _super)
    function MtcnnBox(left, top, right, bottom) {
      return (
        _super.call(
          this,
          { left: left, top: top, right: right, bottom: bottom },
          true
        ) || this
      )
    }
    return MtcnnBox
  })(Box)

  function normalize$1(x) {
    return Ze(function () {
      return gc(Cc(x, On(127.5)), On(0.0078125))
    })
  }

  function prelu(x, alpha) {
    return Ze(function () {
      return rc(Bl(x), gc(alpha, wu(Bl(wu(x)))))
    })
  }

  function sharedLayer(x, params, isPnet) {
    if (isPnet === void 0) {
      isPnet = false
    }
    return Ze(function () {
      var out = convLayer(x, params.conv1, 'valid')
      out = prelu(out, params.prelu1_alpha)
      out = hl(out, isPnet ? [2, 2] : [3, 3], [2, 2], 'same')
      out = convLayer(out, params.conv2, 'valid')
      out = prelu(out, params.prelu2_alpha)
      out = isPnet ? out : hl(out, [3, 3], [2, 2], 'valid')
      out = convLayer(out, params.conv3, 'valid')
      out = prelu(out, params.prelu3_alpha)
      return out
    })
  }

  function PNet(x, params) {
    return Ze(function () {
      var out = sharedLayer(x, params, true)
      var conv = convLayer(out, params.conv4_1, 'valid')
      var max = wr(Sl(conv, 3), 3)
      var prob = go(Cc(conv, max), 3)
      var regions = convLayer(out, params.conv4_2, 'valid')
      return { prob: prob, regions: regions }
    })
  }

  function rescaleAndNormalize(x, scale) {
    return Ze(function () {
      var _a = getSizesForScale(scale, x.shape.slice(1)),
        height = _a.height,
        width = _a.width
      var resized = Oh.resizeBilinear(x, [height, width])
      var normalized = normalize$1(resized)
      return Wl(normalized, [0, 2, 1, 3])
    })
  }
  function extractBoundingBoxes(
    scoresTensor,
    regionsTensor,
    scale,
    scoreThreshold
  ) {
    // TODO: fix this!, maybe better to use tf.gather here
    var indices = []
    var scoresData = scoresTensor.arraySync()
    for (var y = 0; y < scoresTensor.shape[0]; y++) {
      for (var x = 0; x < scoresTensor.shape[1]; x++) {
        if (scoresData[y][x] >= scoreThreshold) {
          indices.push(new Point(x, y))
        }
      }
    }
    var boundingBoxes = indices.map(function (idx) {
      var cell = new BoundingBox(
        Math.round((idx.y * CELL_STRIDE + 1) / scale),
        Math.round((idx.x * CELL_STRIDE + 1) / scale),
        Math.round((idx.y * CELL_STRIDE + CELL_SIZE) / scale),
        Math.round((idx.x * CELL_STRIDE + CELL_SIZE) / scale)
      )
      var score = scoresData[idx.y][idx.x]
      var regionsData = regionsTensor.arraySync()
      var region = new MtcnnBox(
        regionsData[idx.y][idx.x][0],
        regionsData[idx.y][idx.x][1],
        regionsData[idx.y][idx.x][2],
        regionsData[idx.y][idx.x][3]
      )
      return {
        cell: cell,
        score: score,
        region: region,
      }
    })
    return boundingBoxes
  }
  function stage1(imgTensor, scales, scoreThreshold, params, stats) {
    stats.stage1 = []
    var pnetOutputs = scales.map(function (scale) {
      return Ze(function () {
        var statsForScale = { scale: scale }
        var resized = rescaleAndNormalize(imgTensor, scale)
        var ts = Date.now()
        var _a = PNet(resized, params),
          prob = _a.prob,
          regions = _a.regions
        statsForScale.pnet = Date.now() - ts
        var scoresTensor = Ur(Ur(prob, 3)[1])[0]
        var regionsTensor = Ur(regions)[0]
        return {
          scoresTensor: scoresTensor,
          regionsTensor: regionsTensor,
          scale: scale,
          statsForScale: statsForScale,
        }
      })
    })
    var boxesForScale = pnetOutputs.map(function (_a) {
      var scoresTensor = _a.scoresTensor,
        regionsTensor = _a.regionsTensor,
        scale = _a.scale,
        statsForScale = _a.statsForScale
      var boundingBoxes = extractBoundingBoxes(
        scoresTensor,
        regionsTensor,
        scale,
        scoreThreshold
      )
      scoresTensor.dispose()
      regionsTensor.dispose()
      if (!boundingBoxes.length) {
        stats.stage1.push(statsForScale)
        return []
      }
      var ts = Date.now()
      var indices = nonMaxSuppression(
        boundingBoxes.map(function (bbox) {
          return bbox.cell
        }),
        boundingBoxes.map(function (bbox) {
          return bbox.score
        }),
        0.5
      )
      statsForScale.nms = Date.now() - ts
      statsForScale.numBoxes = indices.length
      stats.stage1.push(statsForScale)
      return indices.map(function (boxIdx) {
        return boundingBoxes[boxIdx]
      })
    })
    var allBoxes = boxesForScale.reduce(function (all, boxes) {
      return all.concat(boxes)
    }, [])
    var finalBoxes = []
    var finalScores = []
    if (allBoxes.length > 0) {
      var ts = Date.now()
      var indices = nonMaxSuppression(
        allBoxes.map(function (bbox) {
          return bbox.cell
        }),
        allBoxes.map(function (bbox) {
          return bbox.score
        }),
        0.7
      )
      stats.stage1_nms = Date.now() - ts
      finalScores = indices.map(function (idx) {
        return allBoxes[idx].score
      })
      finalBoxes = indices
        .map(function (idx) {
          return allBoxes[idx]
        })
        .map(function (_a) {
          var cell = _a.cell,
            region = _a.region
          return new BoundingBox(
            cell.left + region.left * cell.width,
            cell.top + region.top * cell.height,
            cell.right + region.right * cell.width,
            cell.bottom + region.bottom * cell.height
          )
            .toSquare()
            .round()
        })
    }
    return {
      boxes: finalBoxes,
      scores: finalScores,
    }
  }

  function extractImagePatches(img, boxes, _a) {
    var width = _a.width,
      height = _a.height
    return __awaiter(this, void 0, void 0, function () {
      var imgCtx, bitmaps, imagePatchesDatas
      var _this = this
      return __generator(this, function (_b) {
        switch (_b.label) {
          case 0:
            imgCtx = getContext2dOrThrow(img)
            return [
              4 /*yield*/,
              Promise.all(
                boxes.map(function (box) {
                  return __awaiter(_this, void 0, void 0, function () {
                    var _a, y, ey, x, ex, fromX, fromY, imgData
                    return __generator(this, function (_b) {
                      ;(_a = box.padAtBorders(img.height, img.width)),
                        (y = _a.y),
                        (ey = _a.ey),
                        (x = _a.x),
                        (ex = _a.ex)
                      fromX = x - 1
                      fromY = y - 1
                      imgData = imgCtx.getImageData(
                        fromX,
                        fromY,
                        ex - fromX,
                        ey - fromY
                      )
                      return [
                        2 /*return*/,
                        env.isNodejs()
                          ? createCanvasFromMedia(imgData)
                          : createImageBitmap(imgData),
                      ]
                    })
                  })
                })
              ),
            ]
          case 1:
            bitmaps = _b.sent()
            imagePatchesDatas = []
            bitmaps.forEach(function (bmp) {
              var patch = createCanvas({ width: width, height: height })
              var patchCtx = getContext2dOrThrow(patch)
              patchCtx.drawImage(bmp, 0, 0, width, height)
              var data = patchCtx.getImageData(0, 0, width, height).data
              var currData = []
              // RGBA -> BGR
              for (var i = 0; i < data.length; i += 4) {
                currData.push(data[i + 2])
                currData.push(data[i + 1])
                currData.push(data[i])
              }
              imagePatchesDatas.push(currData)
            })
            return [
              2 /*return*/,
              imagePatchesDatas.map(function (data) {
                var t = Ze(function () {
                  var imagePatchTensor = Wl(
                    Ln(data, [1, width, height, 3]),
                    [0, 2, 1, 3]
                  ).toFloat()
                  return normalize$1(imagePatchTensor)
                })
                return t
              }),
            ]
        }
      })
    })
  }

  function RNet(x, params) {
    return Ze(function () {
      var convOut = sharedLayer(x, params)
      var vectorized = Or(convOut, [
        convOut.shape[0],
        params.fc1.weights.shape[0],
      ])
      var fc1 = fullyConnectedLayer(vectorized, params.fc1)
      var prelu4 = prelu(fc1, params.prelu4_alpha)
      var fc2_1 = fullyConnectedLayer(prelu4, params.fc2_1)
      var max = wr(Sl(fc2_1, 1), 1)
      var prob = go(Cc(fc2_1, max), 1)
      var regions = fullyConnectedLayer(prelu4, params.fc2_2)
      var scores = Ur(prob, 1)[1]
      return { scores: scores, regions: regions }
    })
  }

  function stage2(img, inputBoxes, scoreThreshold, params, stats) {
    return __awaiter(this, void 0, void 0, function () {
      var ts,
        rnetInputs,
        rnetOuts,
        scoresTensor,
        scores,
        _a,
        _b,
        indices,
        filteredBoxes,
        filteredScores,
        finalBoxes,
        finalScores,
        indicesNms,
        regions_1
      return __generator(this, function (_c) {
        switch (_c.label) {
          case 0:
            ts = Date.now()
            return [
              4 /*yield*/,
              extractImagePatches(img, inputBoxes, { width: 24, height: 24 }),
            ]
          case 1:
            rnetInputs = _c.sent()
            stats.stage2_extractImagePatches = Date.now() - ts
            ts = Date.now()
            rnetOuts = rnetInputs.map(function (rnetInput) {
              var out = RNet(rnetInput, params)
              rnetInput.dispose()
              return out
            })
            stats.stage2_rnet = Date.now() - ts
            scoresTensor =
              rnetOuts.length > 1
                ? Yn(
                    rnetOuts.map(function (out) {
                      return out.scores
                    })
                  )
                : rnetOuts[0].scores
            _b = (_a = Array).from
            return [4 /*yield*/, scoresTensor.data()]
          case 2:
            scores = _b.apply(_a, [_c.sent()])
            scoresTensor.dispose()
            indices = scores
              .map(function (score, idx) {
                return { score: score, idx: idx }
              })
              .filter(function (c) {
                return c.score > scoreThreshold
              })
              .map(function (_a) {
                var idx = _a.idx
                return idx
              })
            filteredBoxes = indices.map(function (idx) {
              return inputBoxes[idx]
            })
            filteredScores = indices.map(function (idx) {
              return scores[idx]
            })
            finalBoxes = []
            finalScores = []
            if (filteredBoxes.length > 0) {
              ts = Date.now()
              indicesNms = nonMaxSuppression(filteredBoxes, filteredScores, 0.7)
              stats.stage2_nms = Date.now() - ts
              regions_1 = indicesNms.map(function (idx) {
                var regionsData = rnetOuts[indices[idx]].regions.arraySync()
                return new MtcnnBox(
                  regionsData[0][0],
                  regionsData[0][1],
                  regionsData[0][2],
                  regionsData[0][3]
                )
              })
              finalScores = indicesNms.map(function (idx) {
                return filteredScores[idx]
              })
              finalBoxes = indicesNms.map(function (idx, i) {
                return filteredBoxes[idx].calibrate(regions_1[i])
              })
            }
            rnetOuts.forEach(function (t) {
              t.regions.dispose()
              t.scores.dispose()
            })
            return [
              2 /*return*/,
              {
                boxes: finalBoxes,
                scores: finalScores,
              },
            ]
        }
      })
    })
  }

  function ONet(x, params) {
    return Ze(function () {
      var out = sharedLayer(x, params)
      out = hl(out, [2, 2], [2, 2], 'same')
      out = convLayer(out, params.conv4, 'valid')
      out = prelu(out, params.prelu4_alpha)
      var vectorized = Or(out, [out.shape[0], params.fc1.weights.shape[0]])
      var fc1 = fullyConnectedLayer(vectorized, params.fc1)
      var prelu5 = prelu(fc1, params.prelu5_alpha)
      var fc2_1 = fullyConnectedLayer(prelu5, params.fc2_1)
      var max = wr(Sl(fc2_1, 1), 1)
      var prob = go(Cc(fc2_1, max), 1)
      var regions = fullyConnectedLayer(prelu5, params.fc2_2)
      var points = fullyConnectedLayer(prelu5, params.fc2_3)
      var scores = Ur(prob, 1)[1]
      return { scores: scores, regions: regions, points: points }
    })
  }

  function stage3(img, inputBoxes, scoreThreshold, params, stats) {
    return __awaiter(this, void 0, void 0, function () {
      var ts,
        onetInputs,
        onetOuts,
        scoresTensor,
        scores,
        _a,
        _b,
        indices,
        filteredRegions,
        filteredBoxes,
        filteredScores,
        finalBoxes,
        finalScores,
        points,
        indicesNms
      return __generator(this, function (_c) {
        switch (_c.label) {
          case 0:
            ts = Date.now()
            return [
              4 /*yield*/,
              extractImagePatches(img, inputBoxes, { width: 48, height: 48 }),
            ]
          case 1:
            onetInputs = _c.sent()
            stats.stage3_extractImagePatches = Date.now() - ts
            ts = Date.now()
            onetOuts = onetInputs.map(function (onetInput) {
              var out = ONet(onetInput, params)
              onetInput.dispose()
              return out
            })
            stats.stage3_onet = Date.now() - ts
            scoresTensor =
              onetOuts.length > 1
                ? Yn(
                    onetOuts.map(function (out) {
                      return out.scores
                    })
                  )
                : onetOuts[0].scores
            _b = (_a = Array).from
            return [4 /*yield*/, scoresTensor.data()]
          case 2:
            scores = _b.apply(_a, [_c.sent()])
            scoresTensor.dispose()
            indices = scores
              .map(function (score, idx) {
                return { score: score, idx: idx }
              })
              .filter(function (c) {
                return c.score > scoreThreshold
              })
              .map(function (_a) {
                var idx = _a.idx
                return idx
              })
            filteredRegions = indices.map(function (idx) {
              var regionsData = onetOuts[idx].regions.arraySync()
              return new MtcnnBox(
                regionsData[0][0],
                regionsData[0][1],
                regionsData[0][2],
                regionsData[0][3]
              )
            })
            filteredBoxes = indices.map(function (idx, i) {
              return inputBoxes[idx].calibrate(filteredRegions[i])
            })
            filteredScores = indices.map(function (idx) {
              return scores[idx]
            })
            finalBoxes = []
            finalScores = []
            points = []
            if (filteredBoxes.length > 0) {
              ts = Date.now()
              indicesNms = nonMaxSuppression(
                filteredBoxes,
                filteredScores,
                0.7,
                false
              )
              stats.stage3_nms = Date.now() - ts
              finalBoxes = indicesNms.map(function (idx) {
                return filteredBoxes[idx]
              })
              finalScores = indicesNms.map(function (idx) {
                return filteredScores[idx]
              })
              points = indicesNms.map(function (idx, i) {
                return Array(5)
                  .fill(0)
                  .map(function (_, ptIdx) {
                    var pointsData = onetOuts[idx].points.arraySync()
                    return new Point(
                      pointsData[0][ptIdx] * (finalBoxes[i].width + 1) +
                        finalBoxes[i].left,
                      pointsData[0][ptIdx + 5] * (finalBoxes[i].height + 1) +
                        finalBoxes[i].top
                    )
                  })
              })
            }
            onetOuts.forEach(function (t) {
              t.regions.dispose()
              t.scores.dispose()
              t.points.dispose()
            })
            return [
              2 /*return*/,
              {
                boxes: finalBoxes,
                scores: finalScores,
                points: points,
              },
            ]
        }
      })
    })
  }

  var Mtcnn = /** @class */ (function (_super) {
    __extends(Mtcnn, _super)
    function Mtcnn() {
      return _super.call(this, 'Mtcnn') || this
    }
    Mtcnn.prototype.load = function (weightsOrUrl) {
      return __awaiter(this, void 0, void 0, function () {
        return __generator(this, function (_a) {
          console.warn('mtcnn is deprecated and will be removed soon')
          return [2 /*return*/, _super.prototype.load.call(this, weightsOrUrl)]
        })
      })
    }
    Mtcnn.prototype.loadFromDisk = function (filePath) {
      return __awaiter(this, void 0, void 0, function () {
        return __generator(this, function (_a) {
          console.warn('mtcnn is deprecated and will be removed soon')
          return [
            2 /*return*/,
            _super.prototype.loadFromDisk.call(this, filePath),
          ]
        })
      })
    }
    Mtcnn.prototype.forwardInput = function (input, forwardParams) {
      if (forwardParams === void 0) {
        forwardParams = {}
      }
      return __awaiter(this, void 0, void 0, function () {
        var params,
          inputCanvas,
          stats,
          tsTotal,
          imgTensor,
          onReturn,
          _a,
          height,
          width,
          _b,
          minFaceSize,
          scaleFactor,
          maxNumScales,
          scoreThresholds,
          scaleSteps,
          scales,
          ts,
          out1,
          out2,
          out3,
          results
        return __generator(this, function (_c) {
          switch (_c.label) {
            case 0:
              params = this.params
              if (!params) {
                throw new Error('Mtcnn - load model before inference')
              }
              inputCanvas = input.canvases[0]
              if (!inputCanvas) {
                throw new Error(
                  'Mtcnn - inputCanvas is not defined, note that passing tensors into Mtcnn.forwardInput is not supported yet.'
                )
              }
              stats = {}
              tsTotal = Date.now()
              imgTensor = Ze(function () {
                return bgrToRgbTensor(wr(id.fromPixels(inputCanvas)).toFloat())
              })
              onReturn = function (results) {
                // dispose tensors on return
                imgTensor.dispose()
                stats.total = Date.now() - tsTotal
                return results
              }
              ;(_a = imgTensor.shape.slice(1)),
                (height = _a[0]),
                (width = _a[1])
              ;(_b = new MtcnnOptions(forwardParams)),
                (minFaceSize = _b.minFaceSize),
                (scaleFactor = _b.scaleFactor),
                (maxNumScales = _b.maxNumScales),
                (scoreThresholds = _b.scoreThresholds),
                (scaleSteps = _b.scaleSteps)
              scales = (
                scaleSteps ||
                pyramidDown(minFaceSize, scaleFactor, [height, width])
              )
                .filter(function (scale) {
                  var sizes = getSizesForScale(scale, [height, width])
                  return Math.min(sizes.width, sizes.height) > CELL_SIZE
                })
                .slice(0, maxNumScales)
              stats.scales = scales
              stats.pyramid = scales.map(function (scale) {
                return getSizesForScale(scale, [height, width])
              })
              ts = Date.now()
              return [
                4 /*yield*/,
                stage1(
                  imgTensor,
                  scales,
                  scoreThresholds[0],
                  params.pnet,
                  stats
                ),
              ]
            case 1:
              out1 = _c.sent()
              stats.total_stage1 = Date.now() - ts
              if (!out1.boxes.length) {
                return [2 /*return*/, onReturn({ results: [], stats: stats })]
              }
              stats.stage2_numInputBoxes = out1.boxes.length
              // using the inputCanvas to extract and resize the image patches, since it is faster
              // than doing this on the gpu
              ts = Date.now()
              return [
                4 /*yield*/,
                stage2(
                  inputCanvas,
                  out1.boxes,
                  scoreThresholds[1],
                  params.rnet,
                  stats
                ),
              ]
            case 2:
              out2 = _c.sent()
              stats.total_stage2 = Date.now() - ts
              if (!out2.boxes.length) {
                return [2 /*return*/, onReturn({ results: [], stats: stats })]
              }
              stats.stage3_numInputBoxes = out2.boxes.length
              ts = Date.now()
              return [
                4 /*yield*/,
                stage3(
                  inputCanvas,
                  out2.boxes,
                  scoreThresholds[2],
                  params.onet,
                  stats
                ),
              ]
            case 3:
              out3 = _c.sent()
              stats.total_stage3 = Date.now() - ts
              results = out3.boxes.map(function (box, idx) {
                return extendWithFaceLandmarks(
                  extendWithFaceDetection(
                    {},
                    new FaceDetection(
                      out3.scores[idx],
                      new Rect(
                        box.left / width,
                        box.top / height,
                        box.width / width,
                        box.height / height
                      ),
                      {
                        height: height,
                        width: width,
                      }
                    )
                  ),
                  new FaceLandmarks5(
                    out3.points[idx].map(function (pt) {
                      return pt
                        .sub(new Point(box.left, box.top))
                        .div(new Point(box.width, box.height))
                    }),
                    { width: box.width, height: box.height }
                  )
                )
              })
              return [
                2 /*return*/,
                onReturn({ results: results, stats: stats }),
              ]
          }
        })
      })
    }
    Mtcnn.prototype.forward = function (input, forwardParams) {
      if (forwardParams === void 0) {
        forwardParams = {}
      }
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [4 /*yield*/, _a.apply(this, [_b.sent(), forwardParams])]
            case 2:
              return [2 /*return*/, _b.sent().results]
          }
        })
      })
    }
    Mtcnn.prototype.forwardWithStats = function (input, forwardParams) {
      if (forwardParams === void 0) {
        forwardParams = {}
      }
      return __awaiter(this, void 0, void 0, function () {
        var _a
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              _a = this.forwardInput
              return [4 /*yield*/, toNetInput(input)]
            case 1:
              return [2 /*return*/, _a.apply(this, [_b.sent(), forwardParams])]
          }
        })
      })
    }
    Mtcnn.prototype.getDefaultModelName = function () {
      return 'mtcnn_model'
    }
    Mtcnn.prototype.extractParamsFromWeigthMap = function (weightMap) {
      return extractParamsFromWeigthMap$7(weightMap)
    }
    Mtcnn.prototype.extractParams = function (weights) {
      return extractParams$7(weights)
    }
    return Mtcnn
  })(NeuralNetwork)

  var IOU_THRESHOLD$1 = 0.4
  var BOX_ANCHORS$1 = [
    new Point(1.603231, 2.094468),
    new Point(6.041143, 7.080126),
    new Point(2.882459, 3.518061),
    new Point(4.266906, 5.178857),
    new Point(9.041765, 10.66308),
  ]
  var MEAN_RGB = [117.001, 114.697, 97.404]

  var TinyFaceDetector = /** @class */ (function (_super) {
    __extends(TinyFaceDetector, _super)
    function TinyFaceDetector() {
      var _this = this
      var config = {
        withSeparableConvs: true,
        iouThreshold: IOU_THRESHOLD$1,
        classes: ['face'],
        anchors: BOX_ANCHORS$1,
        meanRgb: MEAN_RGB,
        isFirstLayerConv2d: true,
        filterSizes: [3, 16, 32, 64, 128, 256, 512],
      }
      _this = _super.call(this, config) || this
      return _this
    }
    Object.defineProperty(TinyFaceDetector.prototype, 'anchors', {
      get: function () {
        return this.config.anchors
      },
      enumerable: true,
      configurable: true,
    })
    TinyFaceDetector.prototype.locateFaces = function (input, forwardParams) {
      return __awaiter(this, void 0, void 0, function () {
        var objectDetections
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.detect(input, forwardParams)]
            case 1:
              objectDetections = _a.sent()
              return [
                2 /*return*/,
                objectDetections.map(function (det) {
                  return new FaceDetection(det.score, det.relativeBox, {
                    width: det.imageWidth,
                    height: det.imageHeight,
                  })
                }),
              ]
          }
        })
      })
    }
    TinyFaceDetector.prototype.getDefaultModelName = function () {
      return 'tiny_face_detector_model'
    }
    TinyFaceDetector.prototype.extractParamsFromWeigthMap = function (
      weightMap
    ) {
      return _super.prototype.extractParamsFromWeigthMap.call(this, weightMap)
    }
    return TinyFaceDetector
  })(TinyYolov2Base)

  var nets = {
    ssdMobilenetv1: new SsdMobilenetv1(),
    tinyFaceDetector: new TinyFaceDetector(),
    tinyYolov2: new TinyYolov2(),
    mtcnn: new Mtcnn(),
    faceLandmark68Net: new FaceLandmark68Net(),
    faceLandmark68TinyNet: new FaceLandmark68TinyNet(),
    faceRecognitionNet: new FaceRecognitionNet(),
    faceExpressionNet: new FaceExpressionNet(),
    ageGenderNet: new AgeGenderNet(),
  }
  /**
   * Attempts to detect all faces in an image using SSD Mobilenetv1 Network.
   *
   * @param input The input image.
   * @param options (optional, default: see SsdMobilenetv1Options constructor for default parameters).
   * @returns Bounding box of each face with score.
   */
  var ssdMobilenetv1 = function (input, options) {
    return nets.ssdMobilenetv1.locateFaces(input, options)
  }
  /**
   * Attempts to detect all faces in an image using the Tiny Face Detector.
   *
   * @param input The input image.
   * @param options (optional, default: see TinyFaceDetectorOptions constructor for default parameters).
   * @returns Bounding box of each face with score.
   */
  var tinyFaceDetector = function (input, options) {
    return nets.tinyFaceDetector.locateFaces(input, options)
  }
  /**
   * Attempts to detect all faces in an image using the Tiny Yolov2 Network.
   *
   * @param input The input image.
   * @param options (optional, default: see TinyYolov2Options constructor for default parameters).
   * @returns Bounding box of each face with score.
   */
  var tinyYolov2 = function (input, options) {
    return nets.tinyYolov2.locateFaces(input, options)
  }
  /**
   * Attempts to detect all faces in an image and the 5 point face landmarks
   * of each detected face using the MTCNN Network.
   *
   * @param input The input image.
   * @param options (optional, default: see MtcnnOptions constructor for default parameters).
   * @returns Bounding box of each face with score and 5 point face landmarks.
   */
  var mtcnn = function (input, options) {
    return nets.mtcnn.forward(input, options)
  }
  /**
   * Detects the 68 point face landmark positions of the face shown in an image.
   *
   * @param inputs The face image extracted from the bounding box of a face. Can
   * also be an array of input images, which will be batch processed.
   * @returns 68 point face landmarks or array thereof in case of batch input.
   */
  var detectFaceLandmarks = function (input) {
    return nets.faceLandmark68Net.detectLandmarks(input)
  }
  /**
   * Detects the 68 point face landmark positions of the face shown in an image
   * using a tinier version of the 68 point face landmark model, which is slightly
   * faster at inference, but also slightly less accurate.
   *
   * @param inputs The face image extracted from the bounding box of a face. Can
   * also be an array of input images, which will be batch processed.
   * @returns 68 point face landmarks or array thereof in case of batch input.
   */
  var detectFaceLandmarksTiny = function (input) {
    return nets.faceLandmark68TinyNet.detectLandmarks(input)
  }
  /**
   * Computes a 128 entry vector (face descriptor / face embeddings) from the face shown in an image,
   * which uniquely represents the features of that persons face. The computed face descriptor can
   * be used to measure the similarity between faces, by computing the euclidean distance of two
   * face descriptors.
   *
   * @param inputs The face image extracted from the aligned bounding box of a face. Can
   * also be an array of input images, which will be batch processed.
   * @returns Face descriptor with 128 entries or array thereof in case of batch input.
   */
  var computeFaceDescriptor = function (input) {
    return nets.faceRecognitionNet.computeFaceDescriptor(input)
  }
  /**
   * Recognizes the facial expressions from a face image.
   *
   * @param inputs The face image extracted from the bounding box of a face. Can
   * also be an array of input images, which will be batch processed.
   * @returns Facial expressions with corresponding probabilities or array thereof in case of batch input.
   */
  var recognizeFaceExpressions = function (input) {
    return nets.faceExpressionNet.predictExpressions(input)
  }
  /**
   * Predicts age and gender from a face image.
   *
   * @param inputs The face image extracted from the bounding box of a face. Can
   * also be an array of input images, which will be batch processed.
   * @returns Predictions with age, gender and gender probability or array thereof in case of batch input.
   */
  var predictAgeAndGender = function (input) {
    return nets.ageGenderNet.predictAgeAndGender(input)
  }
  var loadSsdMobilenetv1Model = function (url) {
    return nets.ssdMobilenetv1.load(url)
  }
  var loadTinyFaceDetectorModel = function (url) {
    return nets.tinyFaceDetector.load(url)
  }
  var loadMtcnnModel = function (url) {
    return nets.mtcnn.load(url)
  }
  var loadTinyYolov2Model = function (url) {
    return nets.tinyYolov2.load(url)
  }
  var loadFaceLandmarkModel = function (url) {
    return nets.faceLandmark68Net.load(url)
  }
  var loadFaceLandmarkTinyModel = function (url) {
    return nets.faceLandmark68TinyNet.load(url)
  }
  var loadFaceRecognitionModel = function (url) {
    return nets.faceRecognitionNet.load(url)
  }
  var loadFaceExpressionModel = function (url) {
    return nets.faceExpressionNet.load(url)
  }
  var loadAgeGenderModel = function (url) {
    return nets.ageGenderNet.load(url)
  }
  // backward compatibility
  var loadFaceDetectionModel = loadSsdMobilenetv1Model
  var locateFaces = ssdMobilenetv1
  var detectLandmarks = detectFaceLandmarks

  var PredictFaceExpressionsTaskBase = /** @class */ (function (_super) {
    __extends(PredictFaceExpressionsTaskBase, _super)
    function PredictFaceExpressionsTaskBase(parentTask, input, extractedFaces) {
      var _this = _super.call(this) || this
      _this.parentTask = parentTask
      _this.input = input
      _this.extractedFaces = extractedFaces
      return _this
    }
    return PredictFaceExpressionsTaskBase
  })(ComposableTask)
  var PredictAllFaceExpressionsTask = /** @class */ (function (_super) {
    __extends(PredictAllFaceExpressionsTask, _super)
    function PredictAllFaceExpressionsTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictAllFaceExpressionsTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResults, faceExpressionsByFace
        var _this = this
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResults = _a.sent()
              return [
                4 /*yield*/,
                extractAllFacesAndComputeResults(
                  parentResults,
                  this.input,
                  function (faces) {
                    return __awaiter(_this, void 0, void 0, function () {
                      return __generator(this, function (_a) {
                        switch (_a.label) {
                          case 0:
                            return [
                              4 /*yield*/,
                              Promise.all(
                                faces.map(function (face) {
                                  return nets.faceExpressionNet.predictExpressions(
                                    face
                                  )
                                })
                              ),
                            ]
                          case 1:
                            return [2 /*return*/, _a.sent()]
                        }
                      })
                    })
                  },
                  this.extractedFaces
                ),
              ]
            case 2:
              faceExpressionsByFace = _a.sent()
              return [
                2 /*return*/,
                parentResults.map(function (parentResult, i) {
                  return extendWithFaceExpressions(
                    parentResult,
                    faceExpressionsByFace[i]
                  )
                }),
              ]
          }
        })
      })
    }
    PredictAllFaceExpressionsTask.prototype.withAgeAndGender = function () {
      return new PredictAllAgeAndGenderTask(this, this.input)
    }
    return PredictAllFaceExpressionsTask
  })(PredictFaceExpressionsTaskBase)
  var PredictSingleFaceExpressionsTask = /** @class */ (function (_super) {
    __extends(PredictSingleFaceExpressionsTask, _super)
    function PredictSingleFaceExpressionsTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictSingleFaceExpressionsTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResult, faceExpressions
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResult = _a.sent()
              if (!parentResult) {
                return [2 /*return*/]
              }
              return [
                4 /*yield*/,
                extractSingleFaceAndComputeResult(
                  parentResult,
                  this.input,
                  function (face) {
                    return nets.faceExpressionNet.predictExpressions(face)
                  },
                  this.extractedFaces
                ),
              ]
            case 2:
              faceExpressions = _a.sent()
              return [
                2 /*return*/,
                extendWithFaceExpressions(parentResult, faceExpressions),
              ]
          }
        })
      })
    }
    PredictSingleFaceExpressionsTask.prototype.withAgeAndGender = function () {
      return new PredictSingleAgeAndGenderTask(this, this.input)
    }
    return PredictSingleFaceExpressionsTask
  })(PredictFaceExpressionsTaskBase)
  var PredictAllFaceExpressionsWithFaceAlignmentTask = /** @class */ (function (
    _super
  ) {
    __extends(PredictAllFaceExpressionsWithFaceAlignmentTask, _super)
    function PredictAllFaceExpressionsWithFaceAlignmentTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictAllFaceExpressionsWithFaceAlignmentTask.prototype.withAgeAndGender =
      function () {
        return new PredictAllAgeAndGenderWithFaceAlignmentTask(this, this.input)
      }
    PredictAllFaceExpressionsWithFaceAlignmentTask.prototype.withFaceDescriptors =
      function () {
        return new ComputeAllFaceDescriptorsTask(this, this.input)
      }
    return PredictAllFaceExpressionsWithFaceAlignmentTask
  })(PredictAllFaceExpressionsTask)
  var PredictSingleFaceExpressionsWithFaceAlignmentTask =
    /** @class */ (function (_super) {
      __extends(PredictSingleFaceExpressionsWithFaceAlignmentTask, _super)
      function PredictSingleFaceExpressionsWithFaceAlignmentTask() {
        return (_super !== null && _super.apply(this, arguments)) || this
      }
      PredictSingleFaceExpressionsWithFaceAlignmentTask.prototype.withAgeAndGender =
        function () {
          return new PredictSingleAgeAndGenderWithFaceAlignmentTask(
            this,
            this.input
          )
        }
      PredictSingleFaceExpressionsWithFaceAlignmentTask.prototype.withFaceDescriptor =
        function () {
          return new ComputeSingleFaceDescriptorTask(this, this.input)
        }
      return PredictSingleFaceExpressionsWithFaceAlignmentTask
    })(PredictSingleFaceExpressionsTask)

  var PredictAgeAndGenderTaskBase = /** @class */ (function (_super) {
    __extends(PredictAgeAndGenderTaskBase, _super)
    function PredictAgeAndGenderTaskBase(parentTask, input, extractedFaces) {
      var _this = _super.call(this) || this
      _this.parentTask = parentTask
      _this.input = input
      _this.extractedFaces = extractedFaces
      return _this
    }
    return PredictAgeAndGenderTaskBase
  })(ComposableTask)
  var PredictAllAgeAndGenderTask = /** @class */ (function (_super) {
    __extends(PredictAllAgeAndGenderTask, _super)
    function PredictAllAgeAndGenderTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictAllAgeAndGenderTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResults, ageAndGenderByFace
        var _this = this
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResults = _a.sent()
              return [
                4 /*yield*/,
                extractAllFacesAndComputeResults(
                  parentResults,
                  this.input,
                  function (faces) {
                    return __awaiter(_this, void 0, void 0, function () {
                      return __generator(this, function (_a) {
                        switch (_a.label) {
                          case 0:
                            return [
                              4 /*yield*/,
                              Promise.all(
                                faces.map(function (face) {
                                  return nets.ageGenderNet.predictAgeAndGender(
                                    face
                                  )
                                })
                              ),
                            ]
                          case 1:
                            return [2 /*return*/, _a.sent()]
                        }
                      })
                    })
                  },
                  this.extractedFaces
                ),
              ]
            case 2:
              ageAndGenderByFace = _a.sent()
              return [
                2 /*return*/,
                parentResults.map(function (parentResult, i) {
                  var _a = ageAndGenderByFace[i],
                    age = _a.age,
                    gender = _a.gender,
                    genderProbability = _a.genderProbability
                  return extendWithAge(
                    extendWithGender(parentResult, gender, genderProbability),
                    age
                  )
                }),
              ]
          }
        })
      })
    }
    PredictAllAgeAndGenderTask.prototype.withFaceExpressions = function () {
      return new PredictAllFaceExpressionsTask(this, this.input)
    }
    return PredictAllAgeAndGenderTask
  })(PredictAgeAndGenderTaskBase)
  var PredictSingleAgeAndGenderTask = /** @class */ (function (_super) {
    __extends(PredictSingleAgeAndGenderTask, _super)
    function PredictSingleAgeAndGenderTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictSingleAgeAndGenderTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResult, _a, age, gender, genderProbability
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResult = _b.sent()
              if (!parentResult) {
                return [2 /*return*/]
              }
              return [
                4 /*yield*/,
                extractSingleFaceAndComputeResult(
                  parentResult,
                  this.input,
                  function (face) {
                    return nets.ageGenderNet.predictAgeAndGender(face)
                  },
                  this.extractedFaces
                ),
              ]
            case 2:
              ;(_a = _b.sent()),
                (age = _a.age),
                (gender = _a.gender),
                (genderProbability = _a.genderProbability)
              return [
                2 /*return*/,
                extendWithAge(
                  extendWithGender(parentResult, gender, genderProbability),
                  age
                ),
              ]
          }
        })
      })
    }
    PredictSingleAgeAndGenderTask.prototype.withFaceExpressions = function () {
      return new PredictSingleFaceExpressionsTask(this, this.input)
    }
    return PredictSingleAgeAndGenderTask
  })(PredictAgeAndGenderTaskBase)
  var PredictAllAgeAndGenderWithFaceAlignmentTask = /** @class */ (function (
    _super
  ) {
    __extends(PredictAllAgeAndGenderWithFaceAlignmentTask, _super)
    function PredictAllAgeAndGenderWithFaceAlignmentTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictAllAgeAndGenderWithFaceAlignmentTask.prototype.withFaceExpressions =
      function () {
        return new PredictAllFaceExpressionsWithFaceAlignmentTask(
          this,
          this.input
        )
      }
    PredictAllAgeAndGenderWithFaceAlignmentTask.prototype.withFaceDescriptors =
      function () {
        return new ComputeAllFaceDescriptorsTask(this, this.input)
      }
    return PredictAllAgeAndGenderWithFaceAlignmentTask
  })(PredictAllAgeAndGenderTask)
  var PredictSingleAgeAndGenderWithFaceAlignmentTask = /** @class */ (function (
    _super
  ) {
    __extends(PredictSingleAgeAndGenderWithFaceAlignmentTask, _super)
    function PredictSingleAgeAndGenderWithFaceAlignmentTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    PredictSingleAgeAndGenderWithFaceAlignmentTask.prototype.withFaceExpressions =
      function () {
        return new PredictSingleFaceExpressionsWithFaceAlignmentTask(
          this,
          this.input
        )
      }
    PredictSingleAgeAndGenderWithFaceAlignmentTask.prototype.withFaceDescriptor =
      function () {
        return new ComputeSingleFaceDescriptorTask(this, this.input)
      }
    return PredictSingleAgeAndGenderWithFaceAlignmentTask
  })(PredictSingleAgeAndGenderTask)

  var ComputeFaceDescriptorsTaskBase = /** @class */ (function (_super) {
    __extends(ComputeFaceDescriptorsTaskBase, _super)
    function ComputeFaceDescriptorsTaskBase(parentTask, input) {
      var _this = _super.call(this) || this
      _this.parentTask = parentTask
      _this.input = input
      return _this
    }
    return ComputeFaceDescriptorsTaskBase
  })(ComposableTask)
  var ComputeAllFaceDescriptorsTask = /** @class */ (function (_super) {
    __extends(ComputeAllFaceDescriptorsTask, _super)
    function ComputeAllFaceDescriptorsTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    ComputeAllFaceDescriptorsTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResults, descriptors
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResults = _a.sent()
              return [
                4 /*yield*/,
                extractAllFacesAndComputeResults(
                  parentResults,
                  this.input,
                  function (faces) {
                    return Promise.all(
                      faces.map(function (face) {
                        return nets.faceRecognitionNet.computeFaceDescriptor(
                          face
                        )
                      })
                    )
                  },
                  null,
                  function (parentResult) {
                    return parentResult.landmarks.align(null, {
                      useDlibAlignment: true,
                    })
                  }
                ),
              ]
            case 2:
              descriptors = _a.sent()
              return [
                2 /*return*/,
                descriptors.map(function (descriptor, i) {
                  return extendWithFaceDescriptor(parentResults[i], descriptor)
                }),
              ]
          }
        })
      })
    }
    ComputeAllFaceDescriptorsTask.prototype.withFaceExpressions = function () {
      return new PredictAllFaceExpressionsWithFaceAlignmentTask(
        this,
        this.input
      )
    }
    ComputeAllFaceDescriptorsTask.prototype.withAgeAndGender = function () {
      return new PredictAllAgeAndGenderWithFaceAlignmentTask(this, this.input)
    }
    return ComputeAllFaceDescriptorsTask
  })(ComputeFaceDescriptorsTaskBase)
  var ComputeSingleFaceDescriptorTask = /** @class */ (function (_super) {
    __extends(ComputeSingleFaceDescriptorTask, _super)
    function ComputeSingleFaceDescriptorTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    ComputeSingleFaceDescriptorTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResult, descriptor
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResult = _a.sent()
              if (!parentResult) {
                return [2 /*return*/]
              }
              return [
                4 /*yield*/,
                extractSingleFaceAndComputeResult(
                  parentResult,
                  this.input,
                  function (face) {
                    return nets.faceRecognitionNet.computeFaceDescriptor(face)
                  },
                  null,
                  function (parentResult) {
                    return parentResult.landmarks.align(null, {
                      useDlibAlignment: true,
                    })
                  }
                ),
              ]
            case 2:
              descriptor = _a.sent()
              return [
                2 /*return*/,
                extendWithFaceDescriptor(parentResult, descriptor),
              ]
          }
        })
      })
    }
    ComputeSingleFaceDescriptorTask.prototype.withFaceExpressions =
      function () {
        return new PredictSingleFaceExpressionsWithFaceAlignmentTask(
          this,
          this.input
        )
      }
    ComputeSingleFaceDescriptorTask.prototype.withAgeAndGender = function () {
      return new PredictSingleAgeAndGenderWithFaceAlignmentTask(
        this,
        this.input
      )
    }
    return ComputeSingleFaceDescriptorTask
  })(ComputeFaceDescriptorsTaskBase)

  var DetectFaceLandmarksTaskBase = /** @class */ (function (_super) {
    __extends(DetectFaceLandmarksTaskBase, _super)
    function DetectFaceLandmarksTaskBase(
      parentTask,
      input,
      useTinyLandmarkNet
    ) {
      var _this = _super.call(this) || this
      _this.parentTask = parentTask
      _this.input = input
      _this.useTinyLandmarkNet = useTinyLandmarkNet
      return _this
    }
    Object.defineProperty(
      DetectFaceLandmarksTaskBase.prototype,
      'landmarkNet',
      {
        get: function () {
          return this.useTinyLandmarkNet
            ? nets.faceLandmark68TinyNet
            : nets.faceLandmark68Net
        },
        enumerable: true,
        configurable: true,
      }
    )
    return DetectFaceLandmarksTaskBase
  })(ComposableTask)
  var DetectAllFaceLandmarksTask = /** @class */ (function (_super) {
    __extends(DetectAllFaceLandmarksTask, _super)
    function DetectAllFaceLandmarksTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    DetectAllFaceLandmarksTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResults, detections, faces, _a, faceLandmarksByFace
        var _this = this
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResults = _b.sent()
              detections = parentResults.map(function (res) {
                return res.detection
              })
              if (!(this.input instanceof wt)) return [3 /*break*/, 3]
              return [4 /*yield*/, extractFaceTensors(this.input, detections)]
            case 2:
              _a = _b.sent()
              return [3 /*break*/, 5]
            case 3:
              return [4 /*yield*/, extractFaces(this.input, detections)]
            case 4:
              _a = _b.sent()
              _b.label = 5
            case 5:
              faces = _a
              return [
                4 /*yield*/,
                Promise.all(
                  faces.map(function (face) {
                    return _this.landmarkNet.detectLandmarks(face)
                  })
                ),
              ]
            case 6:
              faceLandmarksByFace = _b.sent()
              faces.forEach(function (f) {
                return f instanceof wt && f.dispose()
              })
              return [
                2 /*return*/,
                parentResults.map(function (parentResult, i) {
                  return extendWithFaceLandmarks(
                    parentResult,
                    faceLandmarksByFace[i]
                  )
                }),
              ]
          }
        })
      })
    }
    DetectAllFaceLandmarksTask.prototype.withFaceExpressions = function () {
      return new PredictAllFaceExpressionsWithFaceAlignmentTask(
        this,
        this.input
      )
    }
    DetectAllFaceLandmarksTask.prototype.withAgeAndGender = function () {
      return new PredictAllAgeAndGenderWithFaceAlignmentTask(this, this.input)
    }
    DetectAllFaceLandmarksTask.prototype.withFaceDescriptors = function () {
      return new ComputeAllFaceDescriptorsTask(this, this.input)
    }
    return DetectAllFaceLandmarksTask
  })(DetectFaceLandmarksTaskBase)
  var DetectSingleFaceLandmarksTask = /** @class */ (function (_super) {
    __extends(DetectSingleFaceLandmarksTask, _super)
    function DetectSingleFaceLandmarksTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    DetectSingleFaceLandmarksTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var parentResult, detection, faces, _a, landmarks
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              return [4 /*yield*/, this.parentTask]
            case 1:
              parentResult = _b.sent()
              if (!parentResult) {
                return [2 /*return*/]
              }
              detection = parentResult.detection
              if (!(this.input instanceof wt)) return [3 /*break*/, 3]
              return [4 /*yield*/, extractFaceTensors(this.input, [detection])]
            case 2:
              _a = _b.sent()
              return [3 /*break*/, 5]
            case 3:
              return [4 /*yield*/, extractFaces(this.input, [detection])]
            case 4:
              _a = _b.sent()
              _b.label = 5
            case 5:
              faces = _a
              return [4 /*yield*/, this.landmarkNet.detectLandmarks(faces[0])]
            case 6:
              landmarks = _b.sent()
              faces.forEach(function (f) {
                return f instanceof wt && f.dispose()
              })
              return [
                2 /*return*/,
                extendWithFaceLandmarks(parentResult, landmarks),
              ]
          }
        })
      })
    }
    DetectSingleFaceLandmarksTask.prototype.withFaceExpressions = function () {
      return new PredictSingleFaceExpressionsWithFaceAlignmentTask(
        this,
        this.input
      )
    }
    DetectSingleFaceLandmarksTask.prototype.withAgeAndGender = function () {
      return new PredictSingleAgeAndGenderWithFaceAlignmentTask(
        this,
        this.input
      )
    }
    DetectSingleFaceLandmarksTask.prototype.withFaceDescriptor = function () {
      return new ComputeSingleFaceDescriptorTask(this, this.input)
    }
    return DetectSingleFaceLandmarksTask
  })(DetectFaceLandmarksTaskBase)

  var DetectFacesTaskBase = /** @class */ (function (_super) {
    __extends(DetectFacesTaskBase, _super)
    function DetectFacesTaskBase(input, options) {
      if (options === void 0) {
        options = new SsdMobilenetv1Options()
      }
      var _this = _super.call(this) || this
      _this.input = input
      _this.options = options
      return _this
    }
    return DetectFacesTaskBase
  })(ComposableTask)
  var DetectAllFacesTask = /** @class */ (function (_super) {
    __extends(DetectAllFacesTask, _super)
    function DetectAllFacesTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    DetectAllFacesTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var _a, input, options, faceDetectionFunction
        return __generator(this, function (_b) {
          switch (_b.label) {
            case 0:
              ;(_a = this), (input = _a.input), (options = _a.options)
              if (!(options instanceof MtcnnOptions)) return [3 /*break*/, 2]
              return [4 /*yield*/, nets.mtcnn.forward(input, options)]
            case 1:
              return [
                2 /*return*/,
                _b.sent().map(function (result) {
                  return result.detection
                }),
              ]
            case 2:
              faceDetectionFunction =
                options instanceof TinyFaceDetectorOptions
                  ? function (input) {
                      return nets.tinyFaceDetector.locateFaces(input, options)
                    }
                  : options instanceof SsdMobilenetv1Options
                  ? function (input) {
                      return nets.ssdMobilenetv1.locateFaces(input, options)
                    }
                  : options instanceof TinyYolov2Options
                  ? function (input) {
                      return nets.tinyYolov2.locateFaces(input, options)
                    }
                  : null
              if (!faceDetectionFunction) {
                throw new Error(
                  'detectFaces - expected options to be instance of TinyFaceDetectorOptions | SsdMobilenetv1Options | MtcnnOptions | TinyYolov2Options'
                )
              }
              return [2 /*return*/, faceDetectionFunction(input)]
          }
        })
      })
    }
    DetectAllFacesTask.prototype.runAndExtendWithFaceDetections = function () {
      var _this = this
      return new Promise(function (res) {
        return __awaiter(_this, void 0, void 0, function () {
          var detections
          return __generator(this, function (_a) {
            switch (_a.label) {
              case 0:
                return [4 /*yield*/, this.run()]
              case 1:
                detections = _a.sent()
                return [
                  2 /*return*/,
                  res(
                    detections.map(function (detection) {
                      return extendWithFaceDetection({}, detection)
                    })
                  ),
                ]
            }
          })
        })
      })
    }
    DetectAllFacesTask.prototype.withFaceLandmarks = function (
      useTinyLandmarkNet
    ) {
      if (useTinyLandmarkNet === void 0) {
        useTinyLandmarkNet = false
      }
      return new DetectAllFaceLandmarksTask(
        this.runAndExtendWithFaceDetections(),
        this.input,
        useTinyLandmarkNet
      )
    }
    DetectAllFacesTask.prototype.withFaceExpressions = function () {
      return new PredictAllFaceExpressionsTask(
        this.runAndExtendWithFaceDetections(),
        this.input
      )
    }
    DetectAllFacesTask.prototype.withAgeAndGender = function () {
      return new PredictAllAgeAndGenderTask(
        this.runAndExtendWithFaceDetections(),
        this.input
      )
    }
    return DetectAllFacesTask
  })(DetectFacesTaskBase)
  var DetectSingleFaceTask = /** @class */ (function (_super) {
    __extends(DetectSingleFaceTask, _super)
    function DetectSingleFaceTask() {
      return (_super !== null && _super.apply(this, arguments)) || this
    }
    DetectSingleFaceTask.prototype.run = function () {
      return __awaiter(this, void 0, void 0, function () {
        var faceDetections, faceDetectionWithHighestScore
        return __generator(this, function (_a) {
          switch (_a.label) {
            case 0:
              return [
                4 /*yield*/,
                new DetectAllFacesTask(this.input, this.options),
              ]
            case 1:
              faceDetections = _a.sent()
              faceDetectionWithHighestScore = faceDetections[0]
              faceDetections.forEach(function (faceDetection) {
                if (faceDetection.score > faceDetectionWithHighestScore.score) {
                  faceDetectionWithHighestScore = faceDetection
                }
              })
              return [2 /*return*/, faceDetectionWithHighestScore]
          }
        })
      })
    }
    DetectSingleFaceTask.prototype.runAndExtendWithFaceDetection = function () {
      var _this = this
      return new Promise(function (res) {
        return __awaiter(_this, void 0, void 0, function () {
          var detection
          return __generator(this, function (_a) {
            switch (_a.label) {
              case 0:
                return [4 /*yield*/, this.run()]
              case 1:
                detection = _a.sent()
                return [
                  2 /*return*/,
                  res(
                    detection
                      ? extendWithFaceDetection({}, detection)
                      : undefined
                  ),
                ]
            }
          })
        })
      })
    }
    DetectSingleFaceTask.prototype.withFaceLandmarks = function (
      useTinyLandmarkNet
    ) {
      if (useTinyLandmarkNet === void 0) {
        useTinyLandmarkNet = false
      }
      return new DetectSingleFaceLandmarksTask(
        this.runAndExtendWithFaceDetection(),
        this.input,
        useTinyLandmarkNet
      )
    }
    DetectSingleFaceTask.prototype.withFaceExpressions = function () {
      return new PredictSingleFaceExpressionsTask(
        this.runAndExtendWithFaceDetection(),
        this.input
      )
    }
    DetectSingleFaceTask.prototype.withAgeAndGender = function () {
      return new PredictSingleAgeAndGenderTask(
        this.runAndExtendWithFaceDetection(),
        this.input
      )
    }
    return DetectSingleFaceTask
  })(DetectFacesTaskBase)

  function detectSingleFace(input, options) {
    if (options === void 0) {
      options = new SsdMobilenetv1Options()
    }
    return new DetectSingleFaceTask(input, options)
  }
  function detectAllFaces(input, options) {
    if (options === void 0) {
      options = new SsdMobilenetv1Options()
    }
    return new DetectAllFacesTask(input, options)
  }

  // export allFaces API for backward compatibility
  function allFacesSsdMobilenetv1(input, minConfidence) {
    return __awaiter(this, void 0, void 0, function () {
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            console.warn(
              'allFacesSsdMobilenetv1 is deprecated and will be removed soon, use the high level api instead'
            )
            return [
              4 /*yield*/,
              detectAllFaces(
                input,
                new SsdMobilenetv1Options(
                  minConfidence ? { minConfidence: minConfidence } : {}
                )
              )
                .withFaceLandmarks()
                .withFaceDescriptors(),
            ]
          case 1:
            return [2 /*return*/, _a.sent()]
        }
      })
    })
  }
  function allFacesTinyYolov2(input, forwardParams) {
    if (forwardParams === void 0) {
      forwardParams = {}
    }
    return __awaiter(this, void 0, void 0, function () {
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            console.warn(
              'allFacesTinyYolov2 is deprecated and will be removed soon, use the high level api instead'
            )
            return [
              4 /*yield*/,
              detectAllFaces(input, new TinyYolov2Options(forwardParams))
                .withFaceLandmarks()
                .withFaceDescriptors(),
            ]
          case 1:
            return [2 /*return*/, _a.sent()]
        }
      })
    })
  }
  function allFacesMtcnn(input, forwardParams) {
    if (forwardParams === void 0) {
      forwardParams = {}
    }
    return __awaiter(this, void 0, void 0, function () {
      return __generator(this, function (_a) {
        switch (_a.label) {
          case 0:
            console.warn(
              'allFacesMtcnn is deprecated and will be removed soon, use the high level api instead'
            )
            return [
              4 /*yield*/,
              detectAllFaces(input, new MtcnnOptions(forwardParams))
                .withFaceLandmarks()
                .withFaceDescriptors(),
            ]
          case 1:
            return [2 /*return*/, _a.sent()]
        }
      })
    })
  }
  var allFaces = allFacesSsdMobilenetv1

  function euclideanDistance(arr1, arr2) {
    if (arr1.length !== arr2.length)
      throw new Error('euclideanDistance: arr1.length !== arr2.length')
    var desc1 = Array.from(arr1)
    var desc2 = Array.from(arr2)
    return Math.sqrt(
      desc1
        .map(function (val, i) {
          return val - desc2[i]
        })
        .reduce(function (res, diff) {
          return res + Math.pow(diff, 2)
        }, 0)
    )
  }

  var FaceMatcher = /** @class */ (function () {
    function FaceMatcher(inputs, distanceThreshold) {
      if (distanceThreshold === void 0) {
        distanceThreshold = 0.6
      }
      this._distanceThreshold = distanceThreshold
      var inputArray = Array.isArray(inputs) ? inputs : [inputs]
      if (!inputArray.length) {
        throw new Error(
          'FaceRecognizer.constructor - expected atleast one input'
        )
      }
      var count = 1
      var createUniqueLabel = function () {
        return 'person ' + count++
      }
      this._labeledDescriptors = inputArray.map(function (desc) {
        if (desc instanceof LabeledFaceDescriptors) {
          return desc
        }
        if (desc instanceof Float32Array) {
          return new LabeledFaceDescriptors(createUniqueLabel(), [desc])
        }
        if (desc.descriptor && desc.descriptor instanceof Float32Array) {
          return new LabeledFaceDescriptors(createUniqueLabel(), [
            desc.descriptor,
          ])
        }
        throw new Error(
          'FaceRecognizer.constructor - expected inputs to be of type LabeledFaceDescriptors | WithFaceDescriptor<any> | Float32Array | Array<LabeledFaceDescriptors | WithFaceDescriptor<any> | Float32Array>'
        )
      })
    }
    Object.defineProperty(FaceMatcher.prototype, 'labeledDescriptors', {
      get: function () {
        return this._labeledDescriptors
      },
      enumerable: true,
      configurable: true,
    })
    Object.defineProperty(FaceMatcher.prototype, 'distanceThreshold', {
      get: function () {
        return this._distanceThreshold
      },
      enumerable: true,
      configurable: true,
    })
    FaceMatcher.prototype.computeMeanDistance = function (
      queryDescriptor,
      descriptors
    ) {
      return (
        descriptors
          .map(function (d) {
            return euclideanDistance(d, queryDescriptor)
          })
          .reduce(function (d1, d2) {
            return d1 + d2
          }, 0) / (descriptors.length || 1)
      )
    }
    FaceMatcher.prototype.matchDescriptor = function (queryDescriptor) {
      var _this = this
      return this.labeledDescriptors
        .map(function (_a) {
          var descriptors = _a.descriptors,
            label = _a.label
          return new FaceMatch(
            label,
            _this.computeMeanDistance(queryDescriptor, descriptors)
          )
        })
        .reduce(function (best, curr) {
          return best.distance < curr.distance ? best : curr
        })
    }
    FaceMatcher.prototype.findBestMatch = function (queryDescriptor) {
      var bestMatch = this.matchDescriptor(queryDescriptor)
      return bestMatch.distance < this.distanceThreshold
        ? bestMatch
        : new FaceMatch('unknown', bestMatch.distance)
    }
    FaceMatcher.prototype.toJSON = function () {
      return {
        distanceThreshold: this.distanceThreshold,
        labeledDescriptors: this.labeledDescriptors.map(function (ld) {
          return ld.toJSON()
        }),
      }
    }
    FaceMatcher.fromJSON = function (json) {
      var labeledDescriptors = json.labeledDescriptors.map(function (ld) {
        return LabeledFaceDescriptors.fromJSON(ld)
      })
      return new FaceMatcher(labeledDescriptors, json.distanceThreshold)
    }
    return FaceMatcher
  })()

  function createMtcnn(weights) {
    var net = new Mtcnn()
    net.extractWeights(weights)
    return net
  }

  function createTinyFaceDetector(weights) {
    var net = new TinyFaceDetector()
    net.extractWeights(weights)
    return net
  }

  function resizeResults(results, dimensions) {
    var _a = new Dimensions(dimensions.width, dimensions.height),
      width = _a.width,
      height = _a.height
    if (width <= 0 || height <= 0) {
      throw new Error(
        'resizeResults - invalid dimensions: ' +
          JSON.stringify({ width: width, height: height })
      )
    }
    if (Array.isArray(results)) {
      return results.map(function (obj) {
        return resizeResults(obj, { width: width, height: height })
      })
    }
    if (isWithFaceLandmarks(results)) {
      var resizedDetection = results.detection.forSize(width, height)
      var resizedLandmarks = results.unshiftedLandmarks.forSize(
        resizedDetection.box.width,
        resizedDetection.box.height
      )
      return extendWithFaceLandmarks(
        extendWithFaceDetection(results, resizedDetection),
        resizedLandmarks
      )
    }
    if (isWithFaceDetection(results)) {
      return extendWithFaceDetection(
        results,
        results.detection.forSize(width, height)
      )
    }
    if (results instanceof FaceLandmarks || results instanceof FaceDetection) {
      return results.forSize(width, height)
    }
    return results
  }

  exports.AgeGenderNet = AgeGenderNet
  exports.BoundingBox = BoundingBox
  exports.Box = Box
  exports.ComposableTask = ComposableTask
  exports.ComputeAllFaceDescriptorsTask = ComputeAllFaceDescriptorsTask
  exports.ComputeFaceDescriptorsTaskBase = ComputeFaceDescriptorsTaskBase
  exports.ComputeSingleFaceDescriptorTask = ComputeSingleFaceDescriptorTask
  exports.DetectAllFaceLandmarksTask = DetectAllFaceLandmarksTask
  exports.DetectAllFacesTask = DetectAllFacesTask
  exports.DetectFaceLandmarksTaskBase = DetectFaceLandmarksTaskBase
  exports.DetectFacesTaskBase = DetectFacesTaskBase
  exports.DetectSingleFaceLandmarksTask = DetectSingleFaceLandmarksTask
  exports.DetectSingleFaceTask = DetectSingleFaceTask
  exports.Dimensions = Dimensions
  exports.FACE_EXPRESSION_LABELS = FACE_EXPRESSION_LABELS
  exports.FaceDetection = FaceDetection
  exports.FaceDetectionNet = FaceDetectionNet
  exports.FaceExpressionNet = FaceExpressionNet
  exports.FaceExpressions = FaceExpressions
  exports.FaceLandmark68Net = FaceLandmark68Net
  exports.FaceLandmark68TinyNet = FaceLandmark68TinyNet
  exports.FaceLandmarkNet = FaceLandmarkNet
  exports.FaceLandmarks = FaceLandmarks
  exports.FaceLandmarks5 = FaceLandmarks5
  exports.FaceLandmarks68 = FaceLandmarks68
  exports.FaceMatch = FaceMatch
  exports.FaceMatcher = FaceMatcher
  exports.FaceRecognitionNet = FaceRecognitionNet
  exports.LabeledBox = LabeledBox
  exports.LabeledFaceDescriptors = LabeledFaceDescriptors
  exports.Mtcnn = Mtcnn
  exports.MtcnnOptions = MtcnnOptions
  exports.NetInput = NetInput
  exports.NeuralNetwork = NeuralNetwork
  exports.ObjectDetection = ObjectDetection
  exports.Point = Point
  exports.PredictedBox = PredictedBox
  exports.Rect = Rect
  exports.SsdMobilenetv1 = SsdMobilenetv1
  exports.SsdMobilenetv1Options = SsdMobilenetv1Options
  exports.TinyFaceDetector = TinyFaceDetector
  exports.TinyFaceDetectorOptions = TinyFaceDetectorOptions
  exports.TinyYolov2 = TinyYolov2
  exports.TinyYolov2Options = TinyYolov2Options
  exports.allFaces = allFaces
  exports.allFacesMtcnn = allFacesMtcnn
  exports.allFacesSsdMobilenetv1 = allFacesSsdMobilenetv1
  exports.allFacesTinyYolov2 = allFacesTinyYolov2
  exports.awaitMediaLoaded = awaitMediaLoaded
  exports.bufferToImage = bufferToImage
  exports.computeFaceDescriptor = computeFaceDescriptor
  exports.createCanvas = createCanvas
  exports.createCanvasFromMedia = createCanvasFromMedia
  exports.createFaceDetectionNet = createFaceDetectionNet
  exports.createFaceRecognitionNet = createFaceRecognitionNet
  exports.createMtcnn = createMtcnn
  exports.createSsdMobilenetv1 = createSsdMobilenetv1
  exports.createTinyFaceDetector = createTinyFaceDetector
  exports.createTinyYolov2 = createTinyYolov2
  exports.detectAllFaces = detectAllFaces
  exports.detectFaceLandmarks = detectFaceLandmarks
  exports.detectFaceLandmarksTiny = detectFaceLandmarksTiny
  exports.detectLandmarks = detectLandmarks
  exports.detectSingleFace = detectSingleFace
  exports.draw = index$1
  exports.env = env
  exports.euclideanDistance = euclideanDistance
  exports.extendWithAge = extendWithAge
  exports.extendWithFaceDescriptor = extendWithFaceDescriptor
  exports.extendWithFaceDetection = extendWithFaceDetection
  exports.extendWithFaceExpressions = extendWithFaceExpressions
  exports.extendWithFaceLandmarks = extendWithFaceLandmarks
  exports.extendWithGender = extendWithGender
  exports.extractFaceTensors = extractFaceTensors
  exports.extractFaces = extractFaces
  exports.fetchImage = fetchImage
  exports.fetchJson = fetchJson
  exports.fetchNetWeights = fetchNetWeights
  exports.fetchOrThrow = fetchOrThrow
  exports.getContext2dOrThrow = getContext2dOrThrow
  exports.getMediaDimensions = getMediaDimensions
  exports.imageTensorToCanvas = imageTensorToCanvas
  exports.imageToSquare = imageToSquare
  exports.inverseSigmoid = inverseSigmoid
  exports.iou = iou
  exports.isMediaElement = isMediaElement
  exports.isMediaLoaded = isMediaLoaded
  exports.isWithAge = isWithAge
  exports.isWithFaceDetection = isWithFaceDetection
  exports.isWithFaceExpressions = isWithFaceExpressions
  exports.isWithFaceLandmarks = isWithFaceLandmarks
  exports.isWithGender = isWithGender
  exports.loadAgeGenderModel = loadAgeGenderModel
  exports.loadFaceDetectionModel = loadFaceDetectionModel
  exports.loadFaceExpressionModel = loadFaceExpressionModel
  exports.loadFaceLandmarkModel = loadFaceLandmarkModel
  exports.loadFaceLandmarkTinyModel = loadFaceLandmarkTinyModel
  exports.loadFaceRecognitionModel = loadFaceRecognitionModel
  exports.loadMtcnnModel = loadMtcnnModel
  exports.loadSsdMobilenetv1Model = loadSsdMobilenetv1Model
  exports.loadTinyFaceDetectorModel = loadTinyFaceDetectorModel
  exports.loadTinyYolov2Model = loadTinyYolov2Model
  exports.loadWeightMap = loadWeightMap
  exports.locateFaces = locateFaces
  exports.matchDimensions = matchDimensions
  exports.minBbox = minBbox
  exports.mtcnn = mtcnn
  exports.nets = nets
  exports.nonMaxSuppression = nonMaxSuppression
  exports.normalize = normalize
  exports.padToSquare = padToSquare
  exports.predictAgeAndGender = predictAgeAndGender
  exports.recognizeFaceExpressions = recognizeFaceExpressions
  exports.resizeResults = resizeResults
  exports.resolveInput = resolveInput
  exports.shuffleArray = shuffleArray
  exports.sigmoid = sigmoid
  exports.ssdMobilenetv1 = ssdMobilenetv1
  exports.tf = tfCore_esm
  exports.tinyFaceDetector = tinyFaceDetector
  exports.tinyYolov2 = tinyYolov2
  exports.toNetInput = toNetInput
  exports.utils = index
  exports.validateConfig = validateConfig

  Object.defineProperty(exports, '__esModule', { value: true })
})
//# sourceMappingURL=face-api.js.map