Class: DNN::Layers::RNN

Inherits:
Connection show all
Includes:
Activations
Defined in:
lib/dnn/core/rnn_layers.rb

Overview

Super class of all RNN classes.

Direct Known Subclasses

GRU, LSTM, SimpleRNN

Instance Attribute Summary collapse

Attributes inherited from Connection

#l1_lambda, #l2_lambda

Attributes inherited from HasParamLayer

#grads, #params, #trainable

Instance Method Summary collapse

Methods inherited from HasParamLayer

#build, #update

Methods inherited from Layer

#build, #built?, #prev_layer

Constructor Details

#initialize(num_nodes, stateful: false, return_sequences: true, weight_initializer: nil, bias_initializer: nil, l1_lambda: 0, l2_lambda: 0) ⇒ RNN

Returns a new instance of RNN.



12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
 
# File 'lib/dnn/core/rnn_layers.rb', line 12

def initialize(num_nodes,
               stateful: false,
               return_sequences: true,
               weight_initializer: nil,
               bias_initializer: nil,
               l1_lambda: 0,
               l2_lambda: 0)
  super(weight_initializer: weight_initializer, bias_initializer: bias_initializer,
        l1_lambda: l1_lambda, l2_lambda: l2_lambda)
  @num_nodes = num_nodes
  @stateful = stateful
  @return_sequences = return_sequences
  @layers = []
  @h = nil
end

Instance Attribute Details

#h ⇒ Object

Returns the value of attribute h.



8
9
10
 
# File 'lib/dnn/core/rnn_layers.rb', line 8

def h
  @h
end

#num_nodes ⇒ Object (readonly)

Returns the value of attribute num_nodes.



9
10
11
 
# File 'lib/dnn/core/rnn_layers.rb', line 9

def num_nodes
  @num_nodes
end

#stateful ⇒ Object (readonly)

Returns the value of attribute stateful.



10
11
12
 
# File 'lib/dnn/core/rnn_layers.rb', line 10

def stateful
  @stateful
end

Instance Method Details

#backward(dh2s) ⇒ Object 



41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
 
# File 'lib/dnn/core/rnn_layers.rb', line 41

def backward(dh2s)
  @grads[:weight] = Xumo::SFloat.zeros(*@params[:weight].shape)
  @grads[:weight2] = Xumo::SFloat.zeros(*@params[:weight2].shape)
  @grads[:bias] = Xumo::SFloat.zeros(*@params[:bias].shape)
  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
  (0...dh2s.shape[1]).to_a.reverse.each do |t|
    dh2 = dh2s[true, t, false]
    dx, dh = @layers[t].backward(dh2 + dh)
    dxs[true, t, false] = dx
  end
  dxs
end

#dlasso ⇒ Object 



95
96
97
98
99
 
# File 'lib/dnn/core/rnn_layers.rb', line 95

def dlasso
  dlasso = Xumo::SFloat.ones(*@params[:weight].shape)
  dlasso[@params[:weight] < 0] = -1
  @l1_lambda * dlasso
end

#dlasso2 ⇒ Object 



105
106
107
108
109
 
# File 'lib/dnn/core/rnn_layers.rb', line 105

def dlasso2
  dlasso = Xumo::SFloat.ones(*@params[:weight2].shape)
  dlasso[@params[:weight2] < 0] = -1
  @l1_lambda * dlasso
end

#dridge ⇒ Object 



101
102
103
 
# File 'lib/dnn/core/rnn_layers.rb', line 101

def dridge
  @l2_lambda * @params[:weight]
end

#dridge2 ⇒ Object 



111
112
113
 
# File 'lib/dnn/core/rnn_layers.rb', line 111

def dridge2
  @l2_lambda * @params[:weight2]
end

#forward(xs) ⇒ Object 



28
29
30
31
32
33
34
35
36
37
38
39
 
# File 'lib/dnn/core/rnn_layers.rb', line 28

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

#lasso ⇒ Object 



79
80
81
82
83
84
85
 
# File 'lib/dnn/core/rnn_layers.rb', line 79

def lasso
  if @l1_lambda > 0
    @l1_lambda * (@params[:weight].abs.sum + @params[:weight2].abs.sum)
  else
    0
  end
end

#reset_state ⇒ Object 



75
76
77
 
# File 'lib/dnn/core/rnn_layers.rb', line 75

def reset_state
  @h = @h.fill(0) if @h
end

#ridge ⇒ Object 



87
88
89
90
91
92
93
 
# File 'lib/dnn/core/rnn_layers.rb', line 87

def ridge
  if @l2_lambda > 0
    0.5 * (@l2_lambda * ((@params[:weight]**2).sum + (@params[:weight2]**2).sum))
  else
    0
  end
end

#shape ⇒ Object 



71
72
73
 
# File 'lib/dnn/core/rnn_layers.rb', line 71

def shape
  @return_sequences ? [@time_length, @num_nodes] : [@num_nodes]
end

#to_hash(merge_hash = nil) ⇒ Object 



60
61
62
63
64
65
66
67
68
69
 
# File 'lib/dnn/core/rnn_layers.rb', line 60

def to_hash(merge_hash = nil)
  hash = {
    num_nodes: @num_nodes,
    stateful: @stateful,
    return_sequences: @return_sequences,
    h: @h.to_a
  }
  hash.merge!(merge_hash) if merge_hash
  super(hash)
end