Sha256: a4194d2637edf5d689d776e02454953e98a821c0dd3d15ed7ff98ceb77f6ceb1
Contents?: true
Size: 1.5 KB
Versions: 5
Compression:
Stored size: 1.5 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 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
5 entries across 5 versions & 1 rubygems