/
models.py
224 lines (174 loc) · 9.36 KB
/
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
# Model architectures for generators and discriminators
# Torch imports
import pdb
import torch
import torch.nn as nn
import torch.nn.functional as F
#########################################
################2D MODELS###############
#########################################
"""Creates a transposed-convolutional layer, with optional batch normalization."""
def deconv2d(in_channels, out_channels, kernel_size, stride=2, padding=1, output_padding=0, instance_norm=True, reflect_pad=False):
layers = []
layers.append(nn.ConvTranspose2d(in_channels, out_channels, kernel_size, stride, padding, output_padding, bias=False))
if instance_norm:
layers.append(nn.InstanceNorm2d(out_channels))
if reflect_pad:
layers.append(nn.ReflectionPad2d(3))
return nn.Sequential(*layers)
"""Creates a convolutional layer, with optional batch normalization."""
def conv2d(in_channels, out_channels, kernel_size, stride=2, padding=1, instance_norm=True, init_zero_weights=False, reflect_pad=False):
layers = []
if reflect_pad:
layers.append(nn.ReflectionPad2d(3))
conv_layer = nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=kernel_size, stride=stride, padding=padding, bias=False)
if init_zero_weights:
conv_layer.weight.data = torch.randn(out_channels, in_channels, kernel_size, kernel_size) * 0.001
layers.append(conv_layer)
if instance_norm:
layers.append(nn.InstanceNorm2d(out_channels))
return nn.Sequential(*layers)
"""Implementation of ResNet code for preserving input through deep architecture."""
class ResnetBlock2d(nn.Module):
def __init__(self, conv_dim):
super(ResnetBlock2d, self).__init__()
self.conv_layer = conv2d(in_channels=conv_dim, out_channels=conv_dim, kernel_size=3, stride=1, padding=1)
def forward(self, x):
out = x + self.conv_layer(x)
return out
"""Defines the architecture of the generator network (both generators G_XtoY an G_YtoX have the same architecture)."""
class CycleGenerator(nn.Module):
def __init__(self, init_zero_weights=False):
super(CycleGenerator, self).__init__()
#### GENERATOR ARCHITECTURE ####
# 1. Define the encoder part of the generator (that extracts features from the input image)
self.conv1 = conv2d(in_channels=3, out_channels=64, kernel_size=7, stride=1, padding=0, reflect_pad=True)
self.conv2 = conv2d(in_channels=64, out_channels=128, kernel_size=5, stride=2, padding=2) #prev kernel_size = 3, padding = 1
self.conv3 = conv2d(in_channels=128, out_channels=256, kernel_size=3, stride=2, padding=1)
# 2. Define the transformation part of the generator
self.resnet_block1 = ResnetBlock2d(conv_dim=256)
self.resnet_block2 = ResnetBlock2d(conv_dim=256)
self.resnet_block3 = ResnetBlock2d(conv_dim=256)
self.resnet_block4 = ResnetBlock2d(conv_dim=256)
self.resnet_block5 = ResnetBlock2d(conv_dim=256)
self.resnet_block6 = ResnetBlock2d(conv_dim=256)
self.resnet_block7 = ResnetBlock2d(conv_dim=256)
self.resnet_block8 = ResnetBlock2d(conv_dim=256)
self.resnet_block9 = ResnetBlock2d(conv_dim=256)
# 3. Define the decoder part of the generator (that builds up the output image from features)
self.deconv2d_1 = deconv2d(in_channels=256, out_channels=128, kernel_size=3, stride=2, padding=1, output_padding=1)
self.deconv2d_2 = deconv2d(in_channels=128, out_channels=64, kernel_size=5, stride=2, padding=2, output_padding=1) #prev kernel_size = 3, padding = 1
self.conv4 = conv2d(in_channels=64, out_channels=3, kernel_size=7, stride=1, padding=0, reflect_pad=True, instance_norm=False)
def forward(self, x):
"""Generates an image conditioned
on an input image.
Input
-----
x: batch_size x 1 x N x N
Output
------
out: batch_size x 1 x N x N
"""
out = F.relu(self.conv1(x))
out = F.relu(self.conv2(out))
out = F.relu(self.conv3(out))
out = F.relu(self.resnet_block1(out))
out = F.relu(self.resnet_block2(out))
out = F.relu(self.resnet_block3(out))
out = F.relu(self.resnet_block4(out))
out = F.relu(self.resnet_block5(out))
out = F.relu(self.resnet_block6(out))
out = F.relu(self.resnet_block7(out))
out = F.relu(self.resnet_block8(out))
out = F.relu(self.resnet_block9(out))
out = F.relu(self.deconv2d_1(out))
out = F.relu(self.deconv2d_2(out))
out = F.tanh(self.conv4(out))
return out
#XNet encoder
class XNetEncoder(nn.Module):
def __init__(self, init_zero_weights=False):
super(XNetEncoder, self).__init__()
# 1. Define the encoder part of the generator (that extracts features from the input image)
self.conv1 = conv2d(in_channels=3, out_channels=64, kernel_size=7, stride=1, padding=0, reflect_pad=True)
self.conv2 = conv2d(in_channels=64, out_channels=128, kernel_size=3, stride=2, padding=1)
self.conv3 = conv2d(in_channels=128, out_channels=256, kernel_size=3, stride=2, padding=1)
# 2. Define the transformation part of the generator
self.resnet_block1 = ResnetBlock2d(conv_dim=256)
self.resnet_block2 = ResnetBlock2d(conv_dim=256)
self.resnet_block3 = ResnetBlock2d(conv_dim=256)
self.resnet_block4 = ResnetBlock2d(conv_dim=256)
self.resnet_block5 = ResnetBlock2d(conv_dim=256)
self.resnet_block6 = ResnetBlock2d(conv_dim=256)
self.resnet_block7 = ResnetBlock2d(conv_dim=256)
self.resnet_block8 = ResnetBlock2d(conv_dim=256)
self.resnet_block9 = ResnetBlock2d(conv_dim=256)
def forward(self, x):
out = F.relu(self.conv1(x))
out = F.relu(self.conv2(out))
out = F.relu(self.conv3(out))
out = F.relu(self.resnet_block1(out))
out = F.relu(self.resnet_block2(out))
out = F.relu(self.resnet_block3(out))
out = F.relu(self.resnet_block4(out))
out = F.relu(self.resnet_block5(out))
out = F.relu(self.resnet_block6(out))
out = F.relu(self.resnet_block7(out))
out = F.relu(self.resnet_block8(out))
out = F.relu(self.resnet_block9(out))
return out
#XNet decoder
class XNetDecoder(nn.Module):
def __init__(self, init_zero_weights=False):
super(XNetDecoder, self).__init__()
# 3. Define the decoder part of the generator (that builds up the output image from features)
self.deconv2d_1 = deconv2d(in_channels=256, out_channels=128, kernel_size=3, stride=2, padding=1, output_padding=1)
self.deconv2d_2 = deconv2d(in_channels=128, out_channels=64, kernel_size=3, stride=2, padding=1, output_padding=1)
self.conv4 = conv2d(in_channels=64, out_channels=3, kernel_size=7, stride=1, padding=0, reflect_pad=True, instance_norm=False)
def forward(self, x):
out = F.relu(self.deconv2d_1(x))
out = F.relu(self.deconv2d_2(out))
out = F.tanh(self.conv4(out))
return out
#XNet translator
class XNetTranslator(nn.Module):
def __init__(self, init_zero_weights=False):
super(XNetTranslator, self).__init__()
# 2. Define the transformation part of the generator
self.resnet_block1 = ResnetBlock2d(conv_dim=256)
self.resnet_block2 = ResnetBlock2d(conv_dim=256)
self.resnet_block3 = ResnetBlock2d(conv_dim=256)
self.resnet_block4 = ResnetBlock2d(conv_dim=256)
self.resnet_block5 = ResnetBlock2d(conv_dim=256)
self.resnet_block6 = ResnetBlock2d(conv_dim=256)
self.resnet_block7 = ResnetBlock2d(conv_dim=256)
self.resnet_block8 = ResnetBlock2d(conv_dim=256)
self.resnet_block9 = ResnetBlock2d(conv_dim=256)
def forward(self, x):
out = F.relu(self.resnet_block1(x))
out = F.relu(self.resnet_block2(out))
out = F.relu(self.resnet_block3(out))
out = F.relu(self.resnet_block4(out))
out = F.relu(self.resnet_block5(out))
out = F.relu(self.resnet_block6(out))
out = F.relu(self.resnet_block7(out))
out = F.relu(self.resnet_block8(out))
out = F.relu(self.resnet_block9(out))
return out
"""Defines the architecture of the discriminator network (both discriminators D_X and D_Y have the same architecture)."""
class PatchGANDiscriminator(nn.Module):
def __init__(self):
super(PatchGANDiscriminator, self).__init__()
#### ARCHITECTURE ####
self.conv1 = conv2d(in_channels=3, out_channels=64, kernel_size=4, stride=2, padding=1, instance_norm=False)
self.conv2 = conv2d(in_channels=64, out_channels=128, kernel_size=4, stride=2, padding=1)
self.conv3 = conv2d(in_channels=128, out_channels=256, kernel_size=4, stride=2, padding=1)
self.conv4 = conv2d(in_channels=256, out_channels=512, kernel_size=4, stride=2, padding=1)
self.conv5 = conv2d(in_channels=512, out_channels=1, kernel_size=4, stride=1, padding=1, instance_norm=False)
def forward(self, x):
out = F.leaky_relu(self.conv1(x), negative_slope=0.2, inplace=True)
out = F.leaky_relu(self.conv2(out), negative_slope=0.2, inplace=True)
out = F.leaky_relu(self.conv3(out), negative_slope=0.2, inplace=True)
out = F.leaky_relu(self.conv4(out), negative_slope=0.2, inplace=True)
out = F.sigmoid(self.conv5(out))
return out