forked from larryniven/nn
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathlstm-frame.cc
154 lines (120 loc) · 6.07 KB
/
lstm-frame.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
#include "nn/lstm-frame.h"
#include "nn/lstm.h"
#include "nn/nn.h"
#include "nn/lstm-tensor-tree.h"
namespace lstm_frame {
std::shared_ptr<tensor_tree::vertex> make_tensor_tree(int layer)
{
tensor_tree::vertex result { "nil" };
lstm::multilayer_lstm_tensor_tree_factory factory {
std::make_shared<lstm::bi_lstm_tensor_tree_factory>(
lstm::bi_lstm_tensor_tree_factory {
std::make_shared<lstm::lstm_tensor_tree_factory>(
lstm::lstm_tensor_tree_factory{})
}),
layer
};
result.children.push_back(factory());
tensor_tree::vertex fc { "nil" };
fc.children.push_back(tensor_tree::make_tensor("softmax weight"));
fc.children.push_back(tensor_tree::make_tensor("softmax bias"));
result.children.push_back(std::make_shared<tensor_tree::vertex>(fc));
return std::make_shared<tensor_tree::vertex>(result);
}
std::shared_ptr<tensor_tree::vertex> make_dyer_tensor_tree(int layer)
{
tensor_tree::vertex result { "nil" };
lstm::multilayer_lstm_tensor_tree_factory factory {
std::make_shared<lstm::bi_lstm_tensor_tree_factory>(
lstm::bi_lstm_tensor_tree_factory {
std::make_shared<lstm::dyer_lstm_tensor_tree_factory>(
lstm::dyer_lstm_tensor_tree_factory{})
}),
layer
};
result.children.push_back(factory());
tensor_tree::vertex fc { "nil" };
fc.children.push_back(tensor_tree::make_tensor("softmax weight"));
fc.children.push_back(tensor_tree::make_tensor("softmax bias"));
result.children.push_back(std::make_shared<tensor_tree::vertex>(fc));
return std::make_shared<tensor_tree::vertex>(result);
}
std::shared_ptr<lstm::transcriber>
make_transcriber(
std::shared_ptr<tensor_tree::vertex> param,
double dropout,
std::default_random_engine *gen,
bool pyramid)
{
lstm::layered_transcriber result;
int layer = param->children.size();
for (int i = 0; i < layer; ++i) {
std::shared_ptr<lstm::transcriber> f_trans;
std::shared_ptr<lstm::transcriber> b_trans;
f_trans = std::make_shared<lstm::lstm_transcriber>(
lstm::lstm_transcriber { (int) tensor_tree::get_tensor_like(param
->children[i]->children[0]->children[2]).size(0) });
b_trans = std::make_shared<lstm::lstm_transcriber>(
lstm::lstm_transcriber { (int) tensor_tree::get_tensor_like(param
->children[i]->children[1]->children[2]).size(0), true });
if (dropout != 0.0) {
f_trans = std::make_shared<lstm::input_dropout_transcriber>(
lstm::input_dropout_transcriber { f_trans, dropout, *gen });
f_trans = std::make_shared<lstm::output_dropout_transcriber>(
lstm::output_dropout_transcriber { f_trans, dropout, *gen });
b_trans = std::make_shared<lstm::input_dropout_transcriber>(
lstm::input_dropout_transcriber { b_trans, dropout, *gen });
b_trans = std::make_shared<lstm::output_dropout_transcriber>(
lstm::output_dropout_transcriber { b_trans, dropout, *gen });
}
std::shared_ptr<lstm::transcriber> trans = std::make_shared<lstm::bi_transcriber>(
lstm::bi_transcriber { (int) tensor_tree::get_tensor_like(param
->children[i]->children[2]).size(1), f_trans, b_trans });
result.layer.push_back(trans);
if (pyramid && i != layer - 1) {
result.layer.push_back(std::make_shared<lstm::subsampled_transcriber>(
lstm::subsampled_transcriber { 2, 0 }));
}
}
return std::make_shared<lstm::layered_transcriber>(result);
}
std::shared_ptr<lstm::transcriber>
make_dyer_transcriber(
std::shared_ptr<tensor_tree::vertex> param,
double dropout,
std::default_random_engine *gen,
bool pyramid)
{
lstm::layered_transcriber result;
int layer = param->children[0]->children.size();
for (int i = 0; i < layer; ++i) {
std::shared_ptr<lstm::transcriber> f_trans;
std::shared_ptr<lstm::transcriber> b_trans;
f_trans = std::make_shared<lstm::dyer_lstm_transcriber>(
lstm::dyer_lstm_transcriber { (int) tensor_tree::get_tensor_like(param
->children[i]->children[0]->children[2]).size(0) });
b_trans = std::make_shared<lstm::dyer_lstm_transcriber>(
lstm::dyer_lstm_transcriber { (int) tensor_tree::get_tensor_like(param
->children[i]->children[1]->children[2]).size(0), true });
if (dropout != 0.0) {
f_trans = std::make_shared<lstm::input_dropout_transcriber>(
lstm::input_dropout_transcriber { f_trans, dropout, *gen });
f_trans = std::make_shared<lstm::output_dropout_transcriber>(
lstm::output_dropout_transcriber { f_trans, dropout, *gen });
b_trans = std::make_shared<lstm::input_dropout_transcriber>(
lstm::input_dropout_transcriber { b_trans, dropout, *gen });
b_trans = std::make_shared<lstm::output_dropout_transcriber>(
lstm::output_dropout_transcriber { b_trans, dropout, *gen });
}
std::shared_ptr<lstm::transcriber> trans = std::make_shared<lstm::bi_transcriber>(
lstm::bi_transcriber { (int) tensor_tree::get_tensor_like(param
->children[i]->children[2]).size(1), f_trans, b_trans });
result.layer.push_back(trans);
if (pyramid && i != layer - 1) {
result.layer.push_back(std::make_shared<lstm::subsampled_transcriber>(
lstm::subsampled_transcriber { 2, 0 }));
}
}
return std::make_shared<lstm::layered_transcriber>(result);
}
}