-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcifar100.py
145 lines (112 loc) · 5.73 KB
/
cifar100.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
# ===========================================================================
# Project: Sparse Model Soups: A Recipe for Improved Pruning via Model Averaging - IOL Lab @ ZIB
# Paper: arxiv.org/abs/2306.16788
# File: models/cifar100.py
# Description: CIFAR-100 Models
# ===========================================================================
import torch.nn.functional as F
from torch import nn
class WideResNet20(nn.Module):
# WideResNet implementation but with widen_factor=2 and depth=22 instead of 10 and 28 respectively.
def __init__(self, depth=22, widen_factor=10, dropout_rate=0.3, num_classes=100):
super(WideResNet20, self).__init__()
self.in_planes = 16
assert ((depth - 4) % 6 == 0), 'Wide-resnet depth should be 6n+4'
n = (depth - 4) / 6
k = widen_factor
nStages = [16, 16 * k, 32 * k, 64 * k]
class wide_basic(nn.Module):
def __init__(self, in_planes, planes, dropout_rate, stride=1):
super(wide_basic, self).__init__()
self.bn1 = nn.BatchNorm2d(in_planes)
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, padding=1, bias=True)
self.dropout = nn.Dropout(p=dropout_rate)
self.bn2 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=True)
self.shortcut = nn.Sequential()
if stride != 1 or in_planes != planes:
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes, planes, kernel_size=1, stride=stride, bias=True),
)
def forward(self, x):
out = self.dropout(self.conv1(F.relu(self.bn1(x))))
out = self.conv2(F.relu(self.bn2(out)))
out += self.shortcut(x)
return out
self.conv1 = self.conv3x3(3, nStages[0])
self.layer1 = self._wide_layer(wide_basic, nStages[1], n, dropout_rate, stride=1)
self.layer2 = self._wide_layer(wide_basic, nStages[2], n, dropout_rate, stride=2)
self.layer3 = self._wide_layer(wide_basic, nStages[3], n, dropout_rate, stride=2)
self.bn1 = nn.BatchNorm2d(nStages[3], momentum=0.9)
self.linear = nn.Linear(nStages[3], num_classes)
def conv3x3(self, in_planes, out_planes, stride=1):
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=True)
def _wide_layer(self, block, planes, num_blocks, dropout_rate, stride):
strides = [stride] + [1] * (int(num_blocks) - 1)
layers = []
for stride in strides:
layers.append(block(self.in_planes, planes, dropout_rate, stride))
self.in_planes = planes
return nn.Sequential(*layers)
def forward(self, x):
out = self.conv1(x)
out = self.layer1(out)
out = self.layer2(out)
out = self.layer3(out)
out = F.relu(self.bn1(out))
out = F.avg_pool2d(out, 8)
out = out.view(out.size(0), -1)
out = self.linear(out)
return out
class WideResNet(nn.Module):
# Based on https://github.com/meliketoy/wide-resnet.pytorch
def __init__(self, depth=28, widen_factor=10, dropout_rate=0.3, num_classes=100):
super(WideResNet, self).__init__()
self.in_planes = 16
assert ((depth - 4) % 6 == 0), 'Wide-resnet depth should be 6n+4'
n = (depth - 4) / 6
k = widen_factor
nStages = [16, 16 * k, 32 * k, 64 * k]
class wide_basic(nn.Module):
def __init__(self, in_planes, planes, dropout_rate, stride=1):
super(wide_basic, self).__init__()
self.bn1 = nn.BatchNorm2d(in_planes)
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, padding=1, bias=True)
self.dropout = nn.Dropout(p=dropout_rate)
self.bn2 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=True)
self.shortcut = nn.Sequential()
if stride != 1 or in_planes != planes:
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes, planes, kernel_size=1, stride=stride, bias=True),
)
def forward(self, x):
out = self.dropout(self.conv1(F.relu(self.bn1(x))))
out = self.conv2(F.relu(self.bn2(out)))
out += self.shortcut(x)
return out
self.conv1 = self.conv3x3(3, nStages[0])
self.layer1 = self._wide_layer(wide_basic, nStages[1], n, dropout_rate, stride=1)
self.layer2 = self._wide_layer(wide_basic, nStages[2], n, dropout_rate, stride=2)
self.layer3 = self._wide_layer(wide_basic, nStages[3], n, dropout_rate, stride=2)
self.bn1 = nn.BatchNorm2d(nStages[3], momentum=0.9)
self.linear = nn.Linear(nStages[3], num_classes)
def conv3x3(self, in_planes, out_planes, stride=1):
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=True)
def _wide_layer(self, block, planes, num_blocks, dropout_rate, stride):
strides = [stride] + [1] * (int(num_blocks) - 1)
layers = []
for stride in strides:
layers.append(block(self.in_planes, planes, dropout_rate, stride))
self.in_planes = planes
return nn.Sequential(*layers)
def forward(self, x):
out = self.conv1(x)
out = self.layer1(out)
out = self.layer2(out)
out = self.layer3(out)
out = F.relu(self.bn1(out))
out = F.avg_pool2d(out, 8)
out = out.view(out.size(0), -1)
out = self.linear(out)
return out