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
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
|
/*
Copyright (c) 2013, Taiga Nomi
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the <organization> nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#pragma once
#include "tiny_dnn/util/util.h"
#include "tiny_dnn/layers/layer.h"
namespace tiny_dnn {
template<typename Activation>
class partial_connected_layer : public feedforward_layer<Activation> {
public:
CNN_USE_LAYER_MEMBERS;
typedef std::vector<std::pair<serial_size_t, serial_size_t> > io_connections;
typedef std::vector<std::pair<serial_size_t, serial_size_t> > wi_connections;
typedef std::vector<std::pair<serial_size_t, serial_size_t> > wo_connections;
typedef feedforward_layer<Activation> Base;
partial_connected_layer(serial_size_t in_dim,
serial_size_t out_dim,
size_t weight_dim,
size_t bias_dim,
float_t scale_factor = float_t(1))
: Base(std_input_order(bias_dim > 0)),
weight2io_(weight_dim),
out2wi_(out_dim),
in2wo_(in_dim),
bias2out_(bias_dim),
out2bias_(out_dim),
scale_factor_(scale_factor){}
size_t param_size() const {
size_t total_param = 0;
for (auto w : weight2io_)
if (w.size() > 0) total_param++;
for (auto b : bias2out_)
if (b.size() > 0) total_param++;
return total_param;
}
serial_size_t fan_in_size() const override {
return max_size(out2wi_);
}
serial_size_t fan_out_size() const override {
return max_size(in2wo_);
}
void connect_weight(serial_size_t input_index, serial_size_t output_index, serial_size_t weight_index) {
weight2io_[weight_index].emplace_back(input_index, output_index);
out2wi_[output_index].emplace_back(weight_index, input_index);
in2wo_[input_index].emplace_back(weight_index, output_index);
}
void connect_bias(serial_size_t bias_index, serial_size_t output_index) {
out2bias_[output_index] = bias_index;
bias2out_[bias_index].push_back(output_index);
}
void forward_propagation(const std::vector<tensor_t*>& in_data,
std::vector<tensor_t*>& out_data) override {
const tensor_t& in = *in_data[0];
const vec_t& W = (*in_data[1])[0];
const vec_t& b = (*in_data[2])[0];
tensor_t& a = *out_data[1];
// @todo revise the parallelism strategy
for (serial_size_t sample = 0, sample_count = static_cast<serial_size_t>(in.size()); sample < sample_count; ++sample) {
vec_t& a_sample = a[sample];
for_i(parallelize_, out2wi_.size(), [&](int i) {
const wi_connections& connections = out2wi_[i];
float_t& a_element = a_sample[i];
a_element = float_t(0);
for (auto connection : connections)// 13.1%
a_element += W[connection.first] * in[sample][connection.second]; // 3.2%
a_element *= scale_factor_;
a_element += b[out2bias_[i]];
});
}
this->forward_activation(*out_data[0], *out_data[1]);
}
void back_propagation(const std::vector<tensor_t*>& in_data,
const std::vector<tensor_t*>& out_data,
std::vector<tensor_t*>& out_grad,
std::vector<tensor_t*>& in_grad) override {
const tensor_t& prev_out = *in_data[0];
const vec_t& W = (*in_data[1])[0];
vec_t& dW = (*in_grad[1])[0];
vec_t& db = (*in_grad[2])[0];
tensor_t& prev_delta = *in_grad[0];
tensor_t& curr_delta = *out_grad[0];
this->backward_activation(*out_grad[0], *out_data[0], curr_delta);
// @todo revise the parallelism strategy
for (serial_size_t sample = 0, sample_count = static_cast<serial_size_t>(prev_out.size()); sample < sample_count; ++sample) {
for_(parallelize_, 0, in2wo_.size(), [&](const blocked_range& r) {
for (int i = r.begin(); i != r.end(); i++) {
const wo_connections& connections = in2wo_[i];
float_t delta = float_t(0);
for (auto connection : connections)
delta += W[connection.first] * curr_delta[sample][connection.second]; // 40.6%
prev_delta[sample][i] = delta * scale_factor_; // 2.1%
}
});
for_(parallelize_, 0, weight2io_.size(), [&](const blocked_range& r) {
for (int i = r.begin(); i < r.end(); i++) {
const io_connections& connections = weight2io_[i];
float_t diff = float_t(0);
for (auto connection : connections) // 11.9%
diff += prev_out[sample][connection.first] * curr_delta[sample][connection.second];
dW[i] += diff * scale_factor_;
}
});
for (size_t i = 0; i < bias2out_.size(); i++) {
const std::vector<serial_size_t>& outs = bias2out_[i];
float_t diff = float_t(0);
for (auto o : outs)
diff += curr_delta[sample][o];
db[i] += diff;
}
}
}
protected:
std::vector<io_connections> weight2io_; // weight_id -> [(in_id, out_id)]
std::vector<wi_connections> out2wi_; // out_id -> [(weight_id, in_id)]
std::vector<wo_connections> in2wo_; // in_id -> [(weight_id, out_id)]
std::vector<std::vector<serial_size_t> > bias2out_;
std::vector<size_t> out2bias_;
float_t scale_factor_;
};
} // namespace tiny_dnn
|
