-
Notifications
You must be signed in to change notification settings - Fork 2
/
resnet.py
160 lines (128 loc) · 5.31 KB
/
resnet.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
import torch
import torch.nn as nn
import torch.nn.functional as F
# import helpers
class FakeReLU(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
return input.clamp(min=0)
@staticmethod
def backward(ctx, grad_output):
return grad_output
class SequentialWithArgs(torch.nn.Sequential):
def forward(self, input, *args, **kwargs):
vs = list(self._modules.values())
l = len(vs)
for i in range(l):
if i == l-1:
input = vs[i](input, *args, **kwargs)
else:
input = vs[i](input)
return input
class BasicBlock(nn.Module):
expansion = 1
def __init__(self, in_planes, planes, stride=1):
super(BasicBlock, self).__init__()
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(planes)
self.shortcut = nn.Sequential()
if stride != 1 or in_planes != self.expansion*planes:
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes, self.expansion*planes, kernel_size=1,
stride=stride, bias=False),
nn.BatchNorm2d(self.expansion*planes))
def forward(self, x, fake_relu=False):
out = F.relu(self.bn1(self.conv1(x)))
out = self.bn2(self.conv2(out))
out += self.shortcut(x)
if fake_relu:
return FakeReLU.apply(out)
return F.relu(out)
class Bottleneck(nn.Module):
expansion = 4
def __init__(self, in_planes, planes, stride=1):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = nn.Conv2d(planes, self.expansion*planes, kernel_size=1, bias=False)
self.bn3 = nn.BatchNorm2d(self.expansion*planes)
self.shortcut = nn.Sequential()
if stride != 1 or in_planes != self.expansion*planes:
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes, self.expansion*planes, kernel_size=1, stride=stride, bias=False),
nn.BatchNorm2d(self.expansion*planes)
)
def forward(self, x, fake_relu=False):
out = F.relu(self.bn1(self.conv1(x)))
out = F.relu(self.bn2(self.conv2(out)))
out = self.bn3(self.conv3(out))
out += self.shortcut(x)
if fake_relu:
return FakeReLU.apply(out)
return F.relu(out)
class ResNet(nn.Module):
# feat_scale lets us deal with CelebA, other non-32x32 datasets
def __init__(self, block, num_blocks, num_classes=43, feat_scale=1, wm=1):
super(ResNet, self).__init__()
widths = [64, 128, 256, 512]
widths = [int(w * wm) for w in widths]
self.in_planes = widths[0]
self.conv1 = nn.Conv2d(3, self.in_planes, kernel_size=3, stride=1,
padding=1, bias=False)
self.bn1 = nn.BatchNorm2d(self.in_planes)
self.layer1 = self._make_layer(block, widths[0], num_blocks[0], stride=1)
self.layer2 = self._make_layer(block, widths[1], num_blocks[1], stride=2)
self.layer3 = self._make_layer(block, widths[2], num_blocks[2], stride=2)
self.layer4 = self._make_layer(block, widths[3], num_blocks[3], stride=2)
self.linear = nn.Linear(feat_scale*widths[3]*block.expansion, num_classes)
def _make_layer(self, block, planes, num_blocks, stride):
strides = [stride] + [1]*(num_blocks-1)
layers = []
for stride in strides:
layers.append(block(self.in_planes, planes, stride))
self.in_planes = planes * block.expansion
return SequentialWithArgs(*layers)
def forward(self, x, with_latent=False, fake_relu=False):
x = x.type(torch.FloatTensor).to('cuda')
out = F.relu(self.bn1(self.conv1(x)))
out = self.layer1(out)
out = self.layer2(out)
out = self.layer3(out)
out = self.layer4(out, fake_relu=fake_relu)
out = F.avg_pool2d(out, 4)
pre_out = out.view(out.size(0), -1)
final = self.linear(pre_out)
if with_latent:
return final, pre_out
return final
def ResNet18(**kwargs):
return ResNet(BasicBlock, [2,2,2,2], **kwargs)
def ResNet18Wide(**kwargs):
return ResNet(BasicBlock, [2,2,2,2], wd=1.5, **kwargs)
def ResNet18Thin(**kwargs):
return ResNet(BasicBlock, [2,2,2,2], wd=.75, **kwargs)
def ResNet34(**kwargs):
return ResNet(BasicBlock, [3,4,6,3], **kwargs)
def ResNet50(**kwargs):
return ResNet(Bottleneck, [3,4,6,3], **kwargs)
def ResNet101(**kwargs):
return ResNet(Bottleneck, [3,4,23,3], **kwargs)
def ResNet152(**kwargs):
return ResNet(Bottleneck, [3,8,36,3], **kwargs)
resnet50 = ResNet50
resnet18 = ResNet18
resnet101 = ResNet101
resnet152 = ResNet152
# resnet18thin = ResNet18Thin
# resnet18wide = ResNet18Wide
def test():
net = ResNet18()
y = net(torch.randn(1,3,32,32))
print(y.size())