blob: e13ada836e03389f9b36eeaae9061557e2faff73 (
plain)
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
|
/*
* Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
*
* 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
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "ConvolutionSpec.h"
#include "PaddingUtils.h"
#include "ShapeQuery.h"
#include <cassert>
ConvolutionSpec::ConvolutionSpec(const ::caffe::ConvolutionParameter ¶m) : _param(param)
{
// NOTE Dilation is not supported, yet
// TODO Support dilation
assert(param.dilation().size() == 0);
}
uint32_t ConvolutionSpec::group(void) const { return _param.group(); }
uint32_t ConvolutionSpec::channel_axis(void) const
{
return query_on(ifm_shape()).axis(axis_specifier(_param.axis()));
}
uint32_t ConvolutionSpec::pad(uint32_t spatial_axis) const
{
assert(spatial_axis < num_spatial_axes());
auto raw_padding = build_raw_padding().with(_param);
auto spatial_padding = build_spatial_padding(num_spatial_axes()).with(raw_padding);
return spatial_padding.value(spatial_axis);
}
uint32_t ConvolutionSpec::stride(uint32_t spatial_axis) const
{
assert(spatial_axis < num_spatial_axes());
// TODO Support stride_h/stride_w parameters
assert(!_param.has_stride_h());
assert(!_param.has_stride_w());
if (_param.stride().size() == 0)
{
// NOTE default stride is 1
return 1;
}
if (_param.stride().size() == 1)
{
return _param.stride(0);
}
assert(_param.stride().size() == num_spatial_axes());
return _param.stride(spatial_axis);
}
uint32_t ConvolutionSpec::ker_dim(uint32_t spatial_axis) const
{
assert(spatial_axis < num_spatial_axes());
if (_param.kernel_size().size() == 0)
{
if (_param.has_kernel_h() && (spatial_axis == 0))
{
assert(num_spatial_axes() == 2);
return _param.kernel_h();
}
if (_param.has_kernel_w() && (spatial_axis == 1))
{
assert(num_spatial_axes() == 2);
return _param.kernel_w();
}
return 0;
}
assert(!_param.has_kernel_h());
assert(!_param.has_kernel_w());
if (_param.kernel_size().size() == 1)
{
return _param.kernel_size(0);
}
else
{
assert(_param.kernel_size().size() == num_spatial_axes());
return _param.kernel_size(spatial_axis);
}
}
nncc::core::ADT::tensor::Shape ConvolutionSpec::ker_shape(void) const
{
nncc::core::ADT::tensor::Shape res;
res.resize(2 + num_spatial_axes());
res.dim(0) = ker_count();
assert(ifm_dim(channel_axis()) % group() == 0);
res.dim(1) = ifm_dim(channel_axis()) / group();
for (uint32_t axis = 0; axis < num_spatial_axes(); ++axis)
{
res.dim(2 + axis) = ker_dim(axis);
}
return res;
}
nncc::core::ADT::tensor::Shape ConvolutionSpec::ofm_shape(void) const
{
nncc::core::ADT::tensor::Shape res;
res.resize(num_batch_axes() + 1 + num_spatial_axes());
for (uint32_t axis = 0; axis < num_batch_axes(); ++axis)
{
res.dim(axis) = ifm_dim(axis);
}
res.dim(num_batch_axes()) = ker_count();
for (uint32_t spatial_axis = 0; spatial_axis < num_spatial_axes(); ++spatial_axis)
{
const uint32_t full_axis = num_batch_axes() + 1 + spatial_axis;
uint32_t dim = 0;
dim += ifm_dim(full_axis) - ker_dim(spatial_axis) + 2 * pad(spatial_axis);
dim /= stride(spatial_axis);
dim += 1;
res.dim(full_axis) = dim;
}
return res;
}
|
