-
Notifications
You must be signed in to change notification settings - Fork 2
/
resnetv2.py
415 lines (351 loc) · 14.7 KB
/
resnetv2.py
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
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
# Copyright 2020 Huawei Technologies Co., Ltd
#
# 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.
# ============================================================================
"""ResNet."""
import numpy as np
import mindspore.nn as nn
import mindspore.common.dtype as mstype
from mindspore.ops import operations as P
from mindspore.ops import functional as F
from mindspore.common.tensor import Tensor
from scipy.stats import truncnorm
def _conv_variance_scaling_initializer(in_channel, out_channel, kernel_size):
fan_in = in_channel * kernel_size * kernel_size
scale = 1.0
scale /= max(1., fan_in)
stddev = (scale ** 0.5) / .87962566103423978
mu, sigma = 0, stddev
weight = truncnorm(-2, 2, loc=mu, scale=sigma).rvs(out_channel * in_channel * kernel_size * kernel_size)
weight = np.reshape(weight, (out_channel, in_channel, kernel_size, kernel_size))
return Tensor(weight, dtype=mstype.float32)
def _weight_variable(shape, factor=0.01):
init_value = np.random.randn(*shape).astype(np.float32) * factor
return Tensor(init_value)
def _conv3x3(in_channel, out_channel, stride=1, use_se=False):
if use_se:
weight = _conv_variance_scaling_initializer(in_channel, out_channel, kernel_size=3)
else:
weight_shape = (out_channel, in_channel, 3, 3)
weight = _weight_variable(weight_shape)
return nn.Conv2dBnAct(in_channel, out_channel,
kernel_size=3, stride=stride, padding=0, pad_mode='same', weight_init=weight)
def _conv1x1(in_channel, out_channel, stride=1, use_se=False):
if use_se:
weight = _conv_variance_scaling_initializer(in_channel, out_channel, kernel_size=1)
else:
weight_shape = (out_channel, in_channel, 1, 1)
weight = _weight_variable(weight_shape)
return nn.Conv2dBnAct(in_channel, out_channel,
kernel_size=1, stride=stride, padding=0, pad_mode='same', weight_init=weight)
def _conv7x7(in_channel, out_channel, stride=1, use_se=False):
if use_se:
weight = _conv_variance_scaling_initializer(in_channel, out_channel, kernel_size=7)
else:
weight_shape = (out_channel, in_channel, 7, 7)
weight = _weight_variable(weight_shape)
return nn.Conv2dBnAct(in_channel, out_channel,
kernel_size=7, stride=stride, padding=0, pad_mode='same', weight_init=weight)
def _bn(channel):
return nn.BatchNorm2d(channel, eps=1e-4, momentum=0.9,
gamma_init=1, beta_init=0, moving_mean_init=0, moving_var_init=1)
def _bn_last(channel):
return nn.BatchNorm2d(channel, eps=1e-4, momentum=0.9,
gamma_init=0, beta_init=0, moving_mean_init=0, moving_var_init=1)
def _fc(in_channel, out_channel, use_se=False):
if use_se:
weight = np.random.normal(loc=0, scale=0.01, size=out_channel*in_channel)
weight = Tensor(np.reshape(weight, (out_channel, in_channel)), dtype=mstype.float32)
else:
weight_shape = (out_channel, in_channel)
weight = _weight_variable(weight_shape)
return nn.DenseBnAct(in_channel, out_channel, has_bias=True, weight_init=weight, bias_init=0)
class ResidualBlock(nn.Cell):
"""
ResNet V1 residual block definition.
Args:
in_channel (int): Input channel.
out_channel (int): Output channel.
stride (int): Stride size for the first convolutional layer. Default: 1.
use_se (bool): enable SE-ResNet50 net. Default: False.
se_block(bool): use se block in SE-ResNet50 net. Default: False.
Returns:
Tensor, output tensor.
Examples:
>>> ResidualBlock(3, 256, stride=2)
"""
expansion = 4
def __init__(self,
in_channel,
out_channel,
stride=1,
use_se=False, se_block=False):
super(ResidualBlock, self).__init__()
self.stride = stride
self.use_se = use_se
self.se_block = se_block
channel = out_channel // self.expansion
self.conv1 = _conv1x1(in_channel, channel, stride=1, use_se=self.use_se)
self.bn1 = _bn(channel)
if self.use_se and self.stride != 1:
self.e2 = nn.SequentialCell([_conv3x3(channel, channel, stride=1, use_se=True), _bn(channel),
nn.ReLU(), nn.MaxPool2d(kernel_size=2, stride=2, pad_mode='same')])
else:
self.conv2 = _conv3x3(channel, channel, stride=stride, use_se=self.use_se)
self.bn2 = _bn(channel)
self.conv3 = _conv1x1(channel, out_channel, stride=1, use_se=self.use_se)
self.bn3 = _bn_last(out_channel)
if self.se_block:
self.se_global_pool = P.ReduceMean(keep_dims=False)
self.se_dense_0 = _fc(out_channel, int(out_channel/4), use_se=self.use_se)
self.se_dense_1 = _fc(int(out_channel/4), out_channel, use_se=self.use_se)
self.se_sigmoid = nn.Sigmoid()
self.se_mul = P.Mul()
self.relu = nn.ReLU()
self.down_sample = False
if stride != 1 or in_channel != out_channel:
self.down_sample = True
self.down_sample_layer = None
if self.down_sample:
if self.use_se:
if stride == 1:
self.down_sample_layer = nn.SequentialCell([_conv1x1(in_channel, out_channel,
stride, use_se=self.use_se), _bn(out_channel)])
else:
self.down_sample_layer = nn.SequentialCell([nn.MaxPool2d(kernel_size=2, stride=2, pad_mode='same'),
_conv1x1(in_channel, out_channel, 1,
use_se=self.use_se), _bn(out_channel)])
else:
self.down_sample_layer = nn.SequentialCell([_conv1x1(in_channel, out_channel, stride,
use_se=self.use_se), _bn(out_channel)])
self.add = P.TensorAdd()
def construct(self, x):
identity = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
if self.use_se and self.stride != 1:
out = self.e2(out)
else:
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.se_block:
out_se = out
out = self.se_global_pool(out, (2, 3))
out = self.se_dense_0(out)
out = self.relu(out)
out = self.se_dense_1(out)
out = self.se_sigmoid(out)
out = F.reshape(out, F.shape(out) + (1, 1))
out = self.se_mul(out, out_se)
if self.down_sample:
identity = self.down_sample_layer(identity)
out = self.add(out, identity)
out = self.relu(out)
return out
class ResNet(nn.Cell):
"""
ResNet architecture.
Args:
block (Cell): Block for network.
layer_nums (list): Numbers of block in different layers.
in_channels (list): Input channel in each layer.
out_channels (list): Output channel in each layer.
strides (list): Stride size in each layer.
num_classes (int): The number of classes that the training images are belonging to.
use_se (bool): enable SE-ResNet50 net. Default: False.
se_block(bool): use se block in SE-ResNet50 net in layer 3 and layer 4. Default: False.
Returns:
Tensor, output tensor.
Examples:
>>> ResNet(ResidualBlock,
>>> [3, 4, 6, 3],
>>> [64, 256, 512, 1024],
>>> [256, 512, 1024, 2048],
>>> [1, 2, 2, 2],
>>> 10)
"""
def __init__(self,
block,
layer_nums,
in_channels,
out_channels,
strides,
num_classes,
use_se=False):
super(ResNet, self).__init__()
if not len(layer_nums) == len(in_channels) == len(out_channels) == 4:
raise ValueError("the length of layer_num, in_channels, out_channels list must be 4!")
self.use_se = use_se
self.se_block = False
if self.use_se:
self.se_block = True
if self.use_se:
self.conv1_0 = _conv3x3(3, 32, stride=2, use_se=self.use_se)
self.bn1_0 = _bn(32)
self.conv1_1 = _conv3x3(32, 32, stride=1, use_se=self.use_se)
self.bn1_1 = _bn(32)
self.conv1_2 = _conv3x3(32, 64, stride=1, use_se=self.use_se)
else:
self.conv1 = _conv7x7(3, 64, stride=2)
self.bn1 = _bn(64)
self.relu = P.ReLU()
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, pad_mode="same")
self.layer1 = self._make_layer(block,
layer_nums[0],
in_channel=in_channels[0],
out_channel=out_channels[0],
stride=strides[0],
use_se=self.use_se)
self.layer2 = self._make_layer(block,
layer_nums[1],
in_channel=in_channels[1],
out_channel=out_channels[1],
stride=strides[1],
use_se=self.use_se)
self.layer3 = self._make_layer(block,
layer_nums[2],
in_channel=in_channels[2],
out_channel=out_channels[2],
stride=strides[2],
use_se=self.use_se,
se_block=self.se_block)
self.layer4 = self._make_layer(block,
layer_nums[3],
in_channel=in_channels[3],
out_channel=out_channels[3],
stride=strides[3],
use_se=self.use_se,
se_block=self.se_block)
self.mean = P.ReduceMean(keep_dims=True)
self.flatten = nn.Flatten()
self.end_point = _fc(out_channels[3], num_classes, use_se=self.use_se)
def _make_layer(self, block, layer_num, in_channel, out_channel, stride, use_se=False, se_block=False):
"""
Make stage network of ResNet.
Args:
block (Cell): Resnet block.
layer_num (int): Layer number.
in_channel (int): Input channel.
out_channel (int): Output channel.
stride (int): Stride size for the first convolutional layer.
se_block(bool): use se block in SE-ResNet50 net. Default: False.
Returns:
SequentialCell, the output layer.
Examples:
>>> _make_layer(ResidualBlock, 3, 128, 256, 2)
"""
layers = []
resnet_block = block(in_channel, out_channel, stride=stride, use_se=use_se)
layers.append(resnet_block)
if se_block:
for _ in range(1, layer_num - 1):
resnet_block = block(out_channel, out_channel, stride=1, use_se=use_se)
layers.append(resnet_block)
resnet_block = block(out_channel, out_channel, stride=1, use_se=use_se, se_block=se_block)
layers.append(resnet_block)
else:
for _ in range(1, layer_num):
resnet_block = block(out_channel, out_channel, stride=1, use_se=use_se)
layers.append(resnet_block)
return nn.SequentialCell(layers)
def construct(self, x):
if self.use_se:
x = self.conv1_0(x)
x = self.bn1_0(x)
x = self.relu(x)
x = self.conv1_1(x)
x = self.bn1_1(x)
x = self.relu(x)
x = self.conv1_2(x)
else:
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
c1 = self.maxpool(x)
c2 = self.layer1(c1)
c3 = self.layer2(c2)
c4 = self.layer3(c3)
c5 = self.layer4(c4)
out = self.mean(c5, (2, 3))
out = self.flatten(out)
out = self.end_point(out)
return out
def resnet50(class_num=10):
"""
Get ResNet50 neural network.
Args:
class_num (int): Class number.
Returns:
Cell, cell instance of ResNet50 neural network.
Examples:
>>> net = resnet50(10)
"""
return ResNet(ResidualBlock,
[3, 4, 6, 3],
[64, 256, 512, 1024],
[256, 512, 1024, 2048],
[1, 2, 2, 2],
class_num)
def resnet18(class_num=10):
"""
Get ResNet50 neural network.
Args:
class_num (int): Class number.
Returns:
Cell, cell instance of ResNet50 neural network.
Examples:
>>> net = resnet50(10)
"""
return ResNet(ResidualBlock,
[2, 2, 2, 2],
[64,128, 256, 512],
[128, 256, 512, 1024],
[1, 2, 2, 2],
class_num)
def se_resnet50(class_num=1001):
"""
Get SE-ResNet50 neural network.
Args:
class_num (int): Class number.
Returns:
Cell, cell instance of SE-ResNet50 neural network.
Examples:
>>> net = se-resnet50(1001)
"""
return ResNet(ResidualBlock,
[3, 4, 6, 3],
[64, 256, 512, 1024],
[256, 512, 1024, 2048],
[1, 2, 2, 2],
class_num,
use_se=True)
def resnet101(class_num=1001):
"""
Get ResNet101 neural network.
Args:
class_num (int): Class number.
Returns:
Cell, cell instance of ResNet101 neural network.
Examples:
>>> net = resnet101(1001)
"""
return ResNet(ResidualBlock,
[3, 4, 23, 3],
[64, 256, 512, 1024],
[256, 512, 1024, 2048],
[1, 2, 2, 2],
class_num)