Class: DNN::Layers::LSTM

Inherits:
RNN show all
Defined in:
lib/dnn/core/rnn_layers.rb

Instance Attribute Summary

Attributes inherited from RNN

#num_nodes, #return_sequences, #stateful

Attributes inherited from Connection

#bias_initializer, #l1_lambda, #l2_lambda, #weight_initializer

Attributes inherited from HasParamLayer

#params, #trainable

Attributes inherited from Layer

#input_shape

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from RNN

#output_shape, #regularizers, #to_hash

Methods inherited from Connection

#regularizers, #to_hash, #use_bias

Methods inherited from HasParamLayer

#build, #update

Methods inherited from Layer

#build, #built?, #output_shape, #to_hash

Constructor Details

#initialize(num_nodes, stateful: false, return_sequences: true, weight_initializer: RandomNormal.new, bias_initializer: Zeros.new, l1_lambda: 0, l2_lambda: 0, use_bias: true) ⇒ LSTM

Returns a new instance of LSTM.



261
262
263
264
265
266
267
268
269
270
271
 
# File 'lib/dnn/core/rnn_layers.rb', line 261

def initialize(num_nodes,
               stateful: false,
               return_sequences: true,
               weight_initializer: RandomNormal.new,
               bias_initializer: Zeros.new,
               l1_lambda: 0,
               l2_lambda: 0,
               use_bias: true)
  super
  @cell = @params[:c] = Param.new
end

Class Method Details

.load_hash(hash) ⇒ Object 



249
250
251
252
253
254
255
256
257
258
259
 
# File 'lib/dnn/core/rnn_layers.rb', line 249

def self.load_hash(hash)
  lstm = self.new(hash[:num_nodes],
                  stateful: hash[:stateful],
                  return_sequences: hash[:return_sequences],
                  weight_initializer: Utils.load_hash(hash[:weight_initializer]),
                  bias_initializer: Utils.load_hash(hash[:bias_initializer]),
                  l1_lambda: hash[:l1_lambda],
                  l2_lambda: hash[:l2_lambda],
                  use_bias: hash[:use_bias])
  lstm
end

Instance Method Details

#backward(dh2s) ⇒ Object 



294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
 
# File 'lib/dnn/core/rnn_layers.rb', line 294

def backward(dh2s)
  @weight.grad = Xumo::SFloat.zeros(*@weight.data.shape)
  @weight2.grad = Xumo::SFloat.zeros(*@weight2.data.shape)
  @bias.grad = Xumo::SFloat.zeros(*@bias.data.shape) if @bias
  unless @return_sequences
    dh = dh2s
    dh2s = Xumo::SFloat.zeros(dh.shape[0], @time_length, dh.shape[1])
    dh2s[true, -1, false] = dh
  end
  dxs = Xumo::SFloat.zeros(@xs_shape)
  dh = 0
  dc = 0
  (0...dh2s.shape[1]).to_a.reverse.each do |t|
    dh2 = dh2s[true, t, false]
    dx, dh, dc = @layers[t].backward(dh2 + dh, dc)
    dxs[true, t, false] = dx
  end
  dxs
end

#forward(xs) ⇒ Object 



273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
 
# File 'lib/dnn/core/rnn_layers.rb', line 273

def forward(xs)
  @xs_shape = xs.shape
  hs = Xumo::SFloat.zeros(xs.shape[0], @time_length, @num_nodes)
  h = nil
  c = nil
  if @stateful
    h = @hidden.data if @hidden.data
    c = @cell.data if @cell.data
  end
  h ||= Xumo::SFloat.zeros(xs.shape[0], @num_nodes)
  c ||= Xumo::SFloat.zeros(xs.shape[0], @num_nodes)
  xs.shape[1].times do |t|
    x = xs[true, t, false]
    h, c = @layers[t].forward(x, h, c)
    hs[true, t, false] = h
  end
  @hidden.data = h
  @cell.data = c
  @return_sequences ? hs : h
end

#reset_state ⇒ Object 



314
315
316
317
 
# File 'lib/dnn/core/rnn_layers.rb', line 314

def reset_state
  super()
  @cell.data = @cell.data.fill(0) if @cell.data
end