-
Notifications
You must be signed in to change notification settings - Fork 0
/
resnet_models.py
172 lines (139 loc) · 6.14 KB
/
resnet_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
import random
from layers import *
# ------------------- #
# ResNet Example #
# ------------------- #
def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
def conv1x1(in_planes, out_planes, stride=1):
"""1x1 convolution"""
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)
class BasicBlock(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None, groups=1,
base_width=64, dilation=1, norm_layer=None):
super(BasicBlock, self).__init__()
if norm_layer is None:
norm_layer = tdBatchNorm
if groups != 1 or base_width != 64:
raise ValueError('BasicBlock only supports groups=1 and base_width=64')
if dilation > 1:
raise NotImplementedError("Dilation > 1 not supported in BasicBlock")
# Both self.conv1 and self.downsample layers downsample the input when stride != 1
self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = norm_layer(planes)
self.conv2 = conv3x3(planes, planes)
self.bn2 = norm_layer(planes)
self.downsample = downsample
self.stride = stride
self.conv1_s = tdLayer(self.conv1, self.bn1)
self.conv2_s = tdLayer(self.conv2, self.bn2)
self.spike = LIFSpike()
def forward(self, x):
identity = x
out = self.conv1_s(x)
out = self.spike(out)
out = self.conv2_s(out)
if self.downsample is not None:
identity = self.downsample(x)
out += identity
out = self.spike(out)
return out
class ResNet(nn.Module):
def __init__(self, block, layers, num_classes=10, zero_init_residual=False,
groups=1, width_per_group=64, replace_stride_with_dilation=None,
norm_layer=None):
super(ResNet, self).__init__()
if norm_layer is None:
norm_layer = tdBatchNorm
self._norm_layer = norm_layer
self.inplanes = 64
self.dilation = 1
if replace_stride_with_dilation is None:
# each element in the tuple indicates if we should replace
# the 2x2 stride with a dilated convolution instead
replace_stride_with_dilation = [False, False, False]
if len(replace_stride_with_dilation) != 3:
raise ValueError("replace_stride_with_dilation should be None "
"or a 3-element tuple, got {}".format(replace_stride_with_dilation))
self.groups = groups
self.base_width = width_per_group
self.conv1 = nn.Conv2d(3, self.inplanes, kernel_size=3, stride=1, padding=1,
bias=False)
self.bn1 = norm_layer(self.inplanes)
self.conv1_s = tdLayer(self.conv1, self.bn1)
self.layer1 = self._make_layer(block, 128, layers[0])
self.layer2 = self._make_layer(block, 256, layers[1], stride=2,
dilate=replace_stride_with_dilation[0])
self.layer3 = self._make_layer(block, 512, layers[2], stride=2,
dilate=replace_stride_with_dilation[1])
self.avgpool = tdLayer(nn.AdaptiveAvgPool2d((1, 1)))
self.fc1 = nn.Linear(512 * block.expansion, 256)
self.fc1_s = tdLayer(self.fc1)
self.fc2 = nn.Linear(256, num_classes)
self.fc2_s = tdLayer(self.fc2)
self.spike = LIFSpike()
#self.spike_out = LIFSpikeOut()
self.T = 1
# Zero-initialize the last BN in each residual branch,
# so that the residual branch starts with zeros, and each residual block behaves like an identity.
# This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677
if zero_init_residual:
for m in self.modules():
if isinstance(m, Bottleneck):
nn.init.constant_(m.bn3.weight, 0)
elif isinstance(m, BasicBlock):
nn.init.constant_(m.bn2.weight, 0)
elif isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
def _make_layer(self, block, planes, blocks, stride=1, dilate=False):
norm_layer = self._norm_layer
downsample = None
previous_dilation = self.dilation
if dilate:
self.dilation *= stride
stride = 1
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = tdLayer(
conv1x1(self.inplanes, planes * block.expansion, stride),
norm_layer(planes * block.expansion)
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample, self.groups,
self.base_width, previous_dilation, norm_layer))
self.inplanes = planes * block.expansion
for _ in range(1, blocks):
layers.append(block(self.inplanes, planes, groups=self.groups,
base_width=self.base_width, dilation=self.dilation,
norm_layer=norm_layer))
return nn.Sequential(*layers)
def _forward_impl(self, x):
# See note [TorchScript super()]
x = self.conv1_s(x)
x = self.spike(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.avgpool(x)
x = torch.flatten(x, 2)
x = self.fc1_s(x)
x = self.spike(x)
x = self.fc2_s(x)
#x = self.spike_out(x)
return x
def forward(self, x):
x = add_dimention(x, self.T)
return self._forward_impl(x)
def _resnet(block, layers, **kwargs):
model = ResNet(block, layers, **kwargs)
return model
def resnet19(pretrained=False, progress=True, **kwargs):
return _resnet(BasicBlock, [3, 3, 2], **kwargs)
if __name__ == '__main__':
model = resnet19(num_classes=10)
model.T = 3
x = torch.rand(2,3,32,32)
y = model(x)
print(y.shape)