Sha256: 3f8e5e0813e127ebc0212300b2c123f88504b48b43b6608a105a375b02532fff
Contents?: true
Size: 1.97 KB
Versions: 3
Compression:
Stored size: 1.97 KB
Contents
module DNN module Initializers class Initializer def init_param(layer, param_key, param) layer.params[param_key] = param end def to_hash(merge_hash = nil) hash = {class: self.class.name} hash.merge!(merge_hash) if merge_hash hash end end class Zeros < Initializer def init_param(layer, param_key) super(layer, param_key, layer.params[param_key].fill(0)) end end class RandomNormal < Initializer attr_reader :mean attr_reader :std def self.load_hash(hash) self.new(hash[:mean], hash[:std]) end def initialize(mean = 0, std = 0.05) @mean = mean @std = std end def init_param(layer, param_key) super(layer, param_key, layer.params[param_key].rand_norm(@mean, @std)) end def to_hash super({mean: @mean, std: @std}) end end class RandomUniform < Initializer attr_reader :min attr_reader :max def self.load_hash(hash) self.new(hash[:min], hash[:max]) end def initialize(min = -0.25, max = 0.25) @min = min @max = max end def init_param(layer, param_key) super(layer, param_key, layer.params[param_key].rand(@min, @max)) end def to_hash super({min: @min, max: @max}) end end class Xavier < Initializer def init_param(layer, param_key) num_prev_nodes = layer.prev_layer.shape.reduce(:*) super(layer, param_key, layer.params[param_key].rand_norm / Math.sqrt(num_prev_nodes)) end end class He < Initializer def init_param(layer, param_key) num_prev_nodes = layer.prev_layer.shape.reduce(:*) super(layer, param_key, layer.params[param_key].rand_norm / Math.sqrt(num_prev_nodes) * Math.sqrt(2)) end end end end
Version data entries
3 entries across 3 versions & 1 rubygems
Version | Path |
---|---|
ruby-dnn-0.7.1 | lib/dnn/core/initializers.rb |
ruby-dnn-0.7.0 | lib/dnn/core/initializers.rb |
ruby-dnn-0.6.10 | lib/dnn/core/initializers.rb |