-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathunet_models.py
298 lines (232 loc) · 10.6 KB
/
unet_models.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
from torch import nn
from torch.nn import functional as F
import torch
from torchvision import models
import torchvision
def conv3x3(in_, out):
return nn.Conv2d(in_, out, 3, padding=1)
class ConvRelu(nn.Module):
def __init__(self, in_, out):
super().__init__()
self.conv = conv3x3(in_, out)
self.activation = nn.ReLU(inplace=True)
def forward(self, x):
x = self.conv(x)
x = self.activation(x)
return x
class DecoderBlock(nn.Module):
def __init__(self, in_channels, middle_channels, out_channels):
super().__init__()
self.block = nn.Sequential(
ConvRelu(in_channels, middle_channels),
nn.ConvTranspose2d(middle_channels, out_channels, kernel_size=3, stride=2, padding=1, output_padding=1),
nn.ReLU(inplace=True)
)
def forward(self, x):
return self.block(x)
class UNet11(nn.Module):
def __init__(self, num_filters=32, pretrained=False):
"""
:param num_classes:
:param num_filters:
:param pretrained:
False - no pre-trained network is used
True - encoder is pre-trained with VGG11
"""
super().__init__()
self.pool = nn.MaxPool2d(2, 2)
self.encoder = models.vgg11(pretrained=pretrained).features
self.relu = self.encoder[1]
self.conv1 = self.encoder[0]
self.conv2 = self.encoder[3]
self.conv3s = self.encoder[6]
self.conv3 = self.encoder[8]
self.conv4s = self.encoder[11]
self.conv4 = self.encoder[13]
self.conv5s = self.encoder[16]
self.conv5 = self.encoder[18]
self.center = DecoderBlock(num_filters * 8 * 2, num_filters * 8 * 2, num_filters * 8)
self.dec5 = DecoderBlock(num_filters * (16 + 8), num_filters * 8 * 2, num_filters * 8)
self.dec4 = DecoderBlock(num_filters * (16 + 8), num_filters * 8 * 2, num_filters * 4)
self.dec3 = DecoderBlock(num_filters * (8 + 4), num_filters * 4 * 2, num_filters * 2)
self.dec2 = DecoderBlock(num_filters * (4 + 2), num_filters * 2 * 2, num_filters)
self.dec1 = ConvRelu(num_filters * (2 + 1), num_filters)
self.final = nn.Conv2d(num_filters, 1, kernel_size=1)
def forward(self, x):
conv1 = self.relu(self.conv1(x))
conv2 = self.relu(self.conv2(self.pool(conv1)))
conv3s = self.relu(self.conv3s(self.pool(conv2)))
conv3 = self.relu(self.conv3(conv3s))
conv4s = self.relu(self.conv4s(self.pool(conv3)))
conv4 = self.relu(self.conv4(conv4s))
conv5s = self.relu(self.conv5s(self.pool(conv4)))
conv5 = self.relu(self.conv5(conv5s))
center = self.center(self.pool(conv5))
dec5 = self.dec5(torch.cat([center, conv5], 1))
dec4 = self.dec4(torch.cat([dec5, conv4], 1))
dec3 = self.dec3(torch.cat([dec4, conv3], 1))
dec2 = self.dec2(torch.cat([dec3, conv2], 1))
dec1 = self.dec1(torch.cat([dec2, conv1], 1))
return self.final(dec1)
def unet11(pretrained=False, **kwargs):
"""
pretrained:
False - no pre-trained network is used
True - encoder is pre-trained with VGG11
carvana - all weights are pre-trained on
Kaggle: Carvana dataset https://www.kaggle.com/c/carvana-image-masking-challenge
"""
if pretrained== True:
model = UNet11(pretrained=pretrained, **kwargs)
else:
model= UNet11(pretrained=False, **kwargs)
# 根据 pretrained 加载 'carvana'数据集上训练好的网络全部权重
if pretrained == 'carvana':
state = torch.load('TernausNet.pt',map_location={'cuda:0': 'cpu'}) #
model.load_state_dict(state['model']) #,map_location='cpu'
return model
class DecoderBlockV2(nn.Module):
def __init__(self, in_channels, middle_channels, out_channels, is_deconv=True):
super(DecoderBlockV2, self).__init__()
self.in_channels = in_channels
if is_deconv:
"""
Paramaters for Deconvolution were chosen to avoid artifacts, following
link https://distill.pub/2016/deconv-checkerboard/
"""
self.block = nn.Sequential(
ConvRelu(in_channels, middle_channels),
nn.ConvTranspose2d(middle_channels, out_channels, kernel_size=4, stride=2,
padding=1),
nn.ReLU(inplace=True)
)
else:
self.block = nn.Sequential(
nn.Upsample(scale_factor=2, mode='bilinear'),
ConvRelu(in_channels, middle_channels),
ConvRelu(middle_channels, out_channels),
)
def forward(self, x):
return self.block(x)
class AlbuNet(nn.Module):
"""
UNet (https://arxiv.org/abs/1505.04597) with Resnet34(https://arxiv.org/abs/1512.03385) encoder
Proposed by Alexander Buslaev: https://www.linkedin.com/in/al-buslaev/
"""
def __init__(self, num_classes=1, num_filters=32, pretrained=False, is_deconv=False):
"""
:param num_classes:
:param num_filters:
:param pretrained:
False - no pre-trained network is used
True - encoder is pre-trained with resnet34
:is_deconv:
False: bilinear interpolation is used in decoder
True: deconvolution is used in decoder
"""
super().__init__()
self.num_classes = num_classes
self.pool = nn.MaxPool2d(2, 2)
self.encoder = torchvision.models.resnet34(pretrained=pretrained)
self.relu = nn.ReLU(inplace=True)
self.conv1 = nn.Sequential(self.encoder.conv1,
self.encoder.bn1,
self.encoder.relu,
self.pool)
self.conv2 = self.encoder.layer1
self.conv3 = self.encoder.layer2
self.conv4 = self.encoder.layer3
self.conv5 = self.encoder.layer4
self.center = DecoderBlockV2(512, num_filters * 8 * 2, num_filters * 8, is_deconv)
self.dec5 = DecoderBlockV2(512 + num_filters * 8, num_filters * 8 * 2, num_filters * 8, is_deconv)
self.dec4 = DecoderBlockV2(256 + num_filters * 8, num_filters * 8 * 2, num_filters * 8, is_deconv)
self.dec3 = DecoderBlockV2(128 + num_filters * 8, num_filters * 4 * 2, num_filters * 2, is_deconv)
self.dec2 = DecoderBlockV2(64 + num_filters * 2, num_filters * 2 * 2, num_filters * 2 * 2, is_deconv)
self.dec1 = DecoderBlockV2(num_filters * 2 * 2, num_filters * 2 * 2, num_filters, is_deconv)
self.dec0 = ConvRelu(num_filters, num_filters)
self.final = nn.Conv2d(num_filters, num_classes, kernel_size=1)
def forward(self, x):
conv1 = self.conv1(x)
conv2 = self.conv2(conv1)
conv3 = self.conv3(conv2)
conv4 = self.conv4(conv3)
conv5 = self.conv5(conv4)
center = self.center(self.pool(conv5))
dec5 = self.dec5(torch.cat([center, conv5], 1))
dec4 = self.dec4(torch.cat([dec5, conv4], 1))
dec3 = self.dec3(torch.cat([dec4, conv3], 1))
dec2 = self.dec2(torch.cat([dec3, conv2], 1))
dec1 = self.dec1(dec2)
dec0 = self.dec0(dec1)
if self.num_classes > 1:
x_out = F.log_softmax(self.final(dec0), dim=1)
else:
x_out = self.final(dec0)
return x_out
class UNet16(nn.Module):
def __init__(self, num_classes=1, num_filters=32, pretrained=False, is_deconv=False):
"""
:param num_classes:
:param num_filters:
:param pretrained:
False - no pre-trained network used
True - encoder pre-trained with VGG16
:is_deconv:
False: bilinear interpolation is used in decoder
True: deconvolution is used in decoder
"""
super().__init__()
self.num_classes = num_classes
self.pool = nn.MaxPool2d(2, 2)
self.encoder = torchvision.models.vgg16(pretrained=pretrained).features
self.relu = nn.ReLU(inplace=True)
self.conv1 = nn.Sequential(self.encoder[0],
self.relu,
self.encoder[2],
self.relu)
self.conv2 = nn.Sequential(self.encoder[5],
self.relu,
self.encoder[7],
self.relu)
self.conv3 = nn.Sequential(self.encoder[10],
self.relu,
self.encoder[12],
self.relu,
self.encoder[14],
self.relu)
self.conv4 = nn.Sequential(self.encoder[17],
self.relu,
self.encoder[19],
self.relu,
self.encoder[21],
self.relu)
self.conv5 = nn.Sequential(self.encoder[24],
self.relu,
self.encoder[26],
self.relu,
self.encoder[28],
self.relu)
self.center = DecoderBlockV2(512, num_filters * 8 * 2, num_filters * 8, is_deconv)
self.dec5 = DecoderBlockV2(512 + num_filters * 8, num_filters * 8 * 2, num_filters * 8, is_deconv)
self.dec4 = DecoderBlockV2(512 + num_filters * 8, num_filters * 8 * 2, num_filters * 8, is_deconv)
self.dec3 = DecoderBlockV2(256 + num_filters * 8, num_filters * 4 * 2, num_filters * 2, is_deconv)
self.dec2 = DecoderBlockV2(128 + num_filters * 2, num_filters * 2 * 2, num_filters, is_deconv)
self.dec1 = ConvRelu(64 + num_filters, num_filters)
self.final = nn.Conv2d(num_filters, num_classes, kernel_size=1)
def forward(self, x):
conv1 = self.conv1(x)
conv2 = self.conv2(self.pool(conv1))
conv3 = self.conv3(self.pool(conv2))
conv4 = self.conv4(self.pool(conv3))
conv5 = self.conv5(self.pool(conv4))
center = self.center(self.pool(conv5))
dec5 = self.dec5(torch.cat([center, conv5], 1))
dec4 = self.dec4(torch.cat([dec5, conv4], 1))
dec3 = self.dec3(torch.cat([dec4, conv3], 1))
dec2 = self.dec2(torch.cat([dec3, conv2], 1))
dec1 = self.dec1(torch.cat([dec2, conv1], 1))
if self.num_classes > 1:
x_out = F.log_softmax(self.final(dec1), dim=1)
else:
x_out = self.final(dec1)
return x_out