nnet-max-pooling-component.h
6.13 KB
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
174
175
176
// nnet/nnet-max-pooling-component.h
// Copyright 2014 Brno University of Technology (author: Karel Vesely)
// See ../../COPYING for clarification regarding multiple authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
// MERCHANTABLITY OR NON-INFRINGEMENT.
// See the Apache 2 License for the specific language governing permissions and
// limitations under the License.
#ifndef KALDI_NNET_NNET_MAX_POOLING_COMPONENT_H_
#define KALDI_NNET_NNET_MAX_POOLING_COMPONENT_H_
#include <string>
#include <vector>
#include "nnet/nnet-component.h"
#include "nnet/nnet-utils.h"
#include "cudamatrix/cu-math.h"
namespace kaldi {
namespace nnet1 {
/**
* MaxPoolingComponent :
* The input/output matrices are split to submatrices with width 'pool_stride_'.
* The pooling is done over 3rd axis, of the set of 2d matrices.
* Our pooling supports overlaps, overlaps occur when (pool_step_ < pool_size_).
*/
class MaxPoolingComponent : public Component {
public:
MaxPoolingComponent(int32 dim_in, int32 dim_out):
Component(dim_in, dim_out),
pool_size_(0),
pool_step_(0),
pool_stride_(0)
{ }
~MaxPoolingComponent()
{ }
Component* Copy() const { return new MaxPoolingComponent(*this); }
ComponentType GetType() const { return kMaxPoolingComponent; }
void InitData(std::istream &is) {
// parse config
std::string token;
while (is >> std::ws, !is.eof()) {
ReadToken(is, false, &token);
/**/ if (token == "<PoolSize>") ReadBasicType(is, false, &pool_size_);
else if (token == "<PoolStep>") ReadBasicType(is, false, &pool_step_);
else if (token == "<PoolStride>") ReadBasicType(is, false, &pool_stride_);
else KALDI_ERR << "Unknown token " << token << ", a typo in config?"
<< " (PoolSize|PoolStep|PoolStride)";
}
// check
KALDI_ASSERT(pool_size_ != 0 && pool_step_ != 0 && pool_stride_ != 0);
}
void ReadData(std::istream &is, bool binary) {
// pooling hyperparameters
ExpectToken(is, binary, "<PoolSize>");
ReadBasicType(is, binary, &pool_size_);
ExpectToken(is, binary, "<PoolStep>");
ReadBasicType(is, binary, &pool_step_);
ExpectToken(is, binary, "<PoolStride>");
ReadBasicType(is, binary, &pool_stride_);
//
// Sanity checks:
//
// number of patches:
KALDI_ASSERT(input_dim_ % pool_stride_ == 0);
int32 num_patches = input_dim_ / pool_stride_;
// number of pools:
KALDI_ASSERT((num_patches - pool_size_) % pool_step_ == 0);
int32 num_pools = 1 + (num_patches - pool_size_) / pool_step_;
// check output dim:
KALDI_ASSERT(output_dim_ == num_pools * pool_stride_);
//
}
void WriteData(std::ostream &os, bool binary) const {
// pooling hyperparameters
WriteToken(os, binary, "<PoolSize>");
WriteBasicType(os, binary, pool_size_);
WriteToken(os, binary, "<PoolStep>");
WriteBasicType(os, binary, pool_step_);
WriteToken(os, binary, "<PoolStride>");
WriteBasicType(os, binary, pool_stride_);
}
void PropagateFnc(const CuMatrixBase<BaseFloat> &in,
CuMatrixBase<BaseFloat> *out) {
// useful dims
int32 num_patches = input_dim_ / pool_stride_;
int32 num_pools = 1 + (num_patches - pool_size_) / pool_step_;
// do the max-pooling (pools indexed by q)
for (int32 q = 0; q < num_pools; q++) {
// get output buffer of the pool
CuSubMatrix<BaseFloat> pool(out->ColRange(q*pool_stride_, pool_stride_));
pool.Set(-1e20); // reset (large negative value)
for (int32 r = 0; r < pool_size_; r++) { // max
int32 p = r + q * pool_step_; // p = input patch
pool.Max(in.ColRange(p*pool_stride_, pool_stride_));
}
}
}
void BackpropagateFnc(const CuMatrixBase<BaseFloat> &in,
const CuMatrixBase<BaseFloat> &out,
const CuMatrixBase<BaseFloat> &out_diff,
CuMatrixBase<BaseFloat> *in_diff) {
// useful dims
int32 num_patches = input_dim_ / pool_stride_;
int32 num_pools = 1 + (num_patches - pool_size_) / pool_step_;
//
// here we note how many diff matrices are summed for each input patch,
std::vector<int32> patch_summands(num_patches, 0);
// this metainfo will be used to divide diff of patches
// used in more than one pool.
//
in_diff->SetZero(); // reset
for (int32 q = 0; q<num_pools; q++) { // sum
for (int32 r = 0; r<pool_size_; r++) {
int32 p = r + q * pool_step_; // patch number
//
CuSubMatrix<BaseFloat> in_p(in.ColRange(p*pool_stride_, pool_stride_));
CuSubMatrix<BaseFloat> out_q(out.ColRange(q*pool_stride_, pool_stride_));
//
CuSubMatrix<BaseFloat> tgt(in_diff->ColRange(p*pool_stride_, pool_stride_));
CuMatrix<BaseFloat> src(out_diff.ColRange(q*pool_stride_, pool_stride_));
// Only the pool-inputs with 'max-values' are used to back-propagate into,
// the rest of derivatives is zeroed-out by a mask.
CuMatrix<BaseFloat> mask;
in_p.EqualElementMask(out_q, &mask);
src.MulElements(mask);
tgt.AddMat(1.0, src);
patch_summands[p] += 1;
}
}
// divide diff by #summands (compensate for patches used in more pools)
for (int32 p = 0; p < num_patches; p++) {
CuSubMatrix<BaseFloat> tgt(in_diff->ColRange(p*pool_stride_, pool_stride_));
KALDI_ASSERT(patch_summands[p] > 0); // patch at least in one pool
tgt.Scale(1.0/patch_summands[p]);
}
}
private:
int32 pool_size_, // input patches used for pooling
pool_step_, // shift used for pooling (allow overlapping pools)
pool_stride_; // stride used to slice input to a vector of matrices
};
} // namespace nnet1
} // namespace kaldi
#endif // KALDI_NNET_NNET_MAX_POOLING_COMPONENT_H_