-
Notifications
You must be signed in to change notification settings - Fork 7
/
model.py
177 lines (154 loc) · 6.37 KB
/
model.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
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Author : Dengpan Fu ([email protected])
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import numpy as np
import torch
from torch import nn
from torch.nn import functional as F
from torch.nn import init
from torchvision.models import resnet18, resnet34, resnet50
class WideResNetBasicBlock(nn.Module):
def __init__(self, in_planes, out_planes, stride, dropRate=0.0):
super(WideResNetBasicBlock, self).__init__()
self.bn1 = nn.BatchNorm2d(in_planes)
self.relu1 = nn.ReLU(inplace=True)
self.conv1 = nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(out_planes)
self.relu2 = nn.ReLU(inplace=True)
self.conv2 = nn.Conv2d(out_planes, out_planes, kernel_size=3, stride=1,
padding=1, bias=False)
self.droprate = dropRate
self.equalInOut = (in_planes == out_planes)
self.convShortcut = (not self.equalInOut) and nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride,
padding=0, bias=False) or None
def forward(self, x):
if not self.equalInOut:
x = self.relu1(self.bn1(x))
else:
out = self.relu1(self.bn1(x))
out = self.relu2(self.bn2(self.conv1(out if self.equalInOut else x)))
if self.droprate > 0:
out = F.dropout(out, p=self.droprate, training=self.training)
out = self.conv2(out)
return torch.add(x if self.equalInOut else self.convShortcut(x), out)
class WideResNetBlock(nn.Module):
def __init__(self, nb_layers, in_planes, out_planes, block, stride, dropRate=0.0):
super(WideResNetBlock, self).__init__()
self.layer = self._make_layer(block, in_planes, out_planes, nb_layers, stride, dropRate)
def _make_layer(self, block, in_planes, out_planes, nb_layers, stride, dropRate):
layers = []
for i in range(int(nb_layers)):
layers.append(block(i == 0 and in_planes or out_planes, out_planes, i == 0 and stride or 1, dropRate))
return nn.Sequential(*layers)
def forward(self, x):
return self.layer(x)
class WideResNet(nn.Module):
def __init__(self, depth=16, num_classes=10, widen_factor=10, dropRate=0.0):
super(WideResNet, self).__init__()
nChannels = [16, 16*widen_factor, 32*widen_factor, 64*widen_factor]
assert((depth - 4) % 6 == 0)
n = (depth - 4) / 6
block = WideResNetBasicBlock
# 1st conv before any network block
self.conv1 = nn.Conv2d(3, nChannels[0], kernel_size=3, stride=1,
padding=1, bias=False)
# 1st block
self.block1 = WideResNetBlock(n, nChannels[0], nChannels[1], block, 1, dropRate)
# 2nd block
self.block2 = WideResNetBlock(n, nChannels[1], nChannels[2], block, 2, dropRate)
# 3rd block
self.block3 = WideResNetBlock(n, nChannels[2], nChannels[3], block, 2, dropRate)
# global average pooling and classifier
self.bn1 = nn.BatchNorm2d(nChannels[3])
self.relu = nn.ReLU(inplace=True)
self.fc = nn.Linear(nChannels[3], num_classes)
self.nChannels = nChannels[3]
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, np.sqrt(2. / n))
# init.kaiming_normal_(m.weight)
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
elif isinstance(m, nn.Linear):
tmp = np.sqrt(3. / m.weight.data.shape[0])
m.weight.data.uniform_(-tmp, tmp)
m.bias.data.zero_()
# init.kaiming_normal_(m.weight)
# init.constant_(m.bias, 0)
def forward(self, x):
out = self.conv1(x)
out = self.block1(out)
out = self.block2(out)
out = self.block3(out)
out = self.relu(self.bn1(out))
out = F.avg_pool2d(out, 8)
out = out.view(-1, self.nChannels)
return self.fc(out)
def wideresnet16(**kwargs):
return WideResNet(depth=16, **kwargs)
def wideresnet22(**kwargs):
return WideResNet(depth=22, **kwargs)
class MnistModel(nn.Module):
""" Construct basic MnistModel for mnist adversal attack """
def __init__(self, re_init=False, has_dropout=False):
super(MnistModel, self).__init__()
self.re_init = re_init
self.has_dropout = has_dropout
self.conv1 = nn.Conv2d(1, 32, kernel_size=5, stride=1, padding=2)
self.conv2 = nn.Conv2d(32, 64, kernel_size=5, stride=1, padding=2)
self.pool = nn.MaxPool2d(2)
self.relu = nn.ReLU(True)
self.fc1 = nn.Linear(7*7*64, 1024)
self.fc2 = nn.Linear(1024, 10)
if self.has_dropout:
self.dropout = nn.Dropout()
if self.re_init:
self._init_params(self.conv1)
self._init_params(self.conv2)
self._init_params(self.fc1)
self._init_params(self.fc2)
def forward(self, x):
x = self.conv1(x)
x = self.relu(x)
x = self.pool(x)
x = self.conv2(x)
x = self.relu(x)
x = self.pool(x)
x = x.view(x.size(0), -1)
x = self.fc1(x)
x = self.relu(x)
if self.has_dropout:
x = self.dropout(x)
x = self.fc2(x)
return x
def _init_params(self, module, mean=0.1, std=0.1):
init.normal_(module.weight, std=0.1)
if hasattr(module, 'bias'):
init.constant_(module.bias, mean)
__factory = {
# resnet series, kwargs: num_classes
'resnet': resnet18,
'resnet18': resnet18,
'resnet34': resnet34,
'resnet50': resnet50,
# wideresnet series, kwargs: num_classes, widen_factor, dropRate
'wide': wideresnet16,
'wideresnet': wideresnet16,
'wideresnet16': wideresnet16,
'wideresnet22': wideresnet22,
# mnist, kwargs: has_dropout
'mnist': MnistModel,
}
def create_model(name, **kwargs):
assert(name in __factory), 'invalid network'
return __factory[name](**kwargs)
if __name__ == '__main__':
net = create_model('wide')
import pdb; pdb.set_trace() # breakpoint 2e2204d9 //