-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCIFAR-100_training.py
204 lines (174 loc) · 8.73 KB
/
CIFAR-100_training.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
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Fri Apr 17 10:13:27 2020
@author: lee
"""
import torch
from torchvision import datasets
import torch.nn as nn
from torchvision import transforms
import argparse
from torch.utils.data import DataLoader
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from models.wresnet import *
from torch.nn.parallel.data_parallel import DataParallel
import torch.backends.cudnn as cudnn
class DataParallelCriterion(DataParallel):
def forward(self, inputs, *targets, **kwargs):
targets, kwargs = self.scatter(targets, kwargs, self.device_ids)
replicas = self.replicate(self.module, self.device_ids[:len(inputs)])
targets = tuple(targets_per_gpu[0] for targets_per_gpu in targets)
outputs = _criterion_parallel_apply(replicas, inputs, targets, kwargs)
return Reduce.apply(*outputs) / len(outputs), targets
def accuracy(output, target, topk=(1,)):
"""Computes the precision@k for the specified values of k"""
if len(target.shape) > 1:
target = torch.argmax(target, dim=1)
maxk = max(topk)
batch_size = target.size(0)
_, pred = output.topk(maxk, 1, True, True)
pred = pred.t()
correct = pred.eq(target.view(1, -1).expand_as(pred))
res = []
for k in topk:
correct_k = correct[:k].view(-1).float().sum(0)
res.append(correct_k.mul_(1 / batch_size))
return res
class BasicBlock(nn.Module):
def __init__(self, in_planes, out_planes, stride, dropRate=0.0):
super(BasicBlock, 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 NetworkBlock(nn.Module):
def __init__(self, nb_layers, in_planes, out_planes, block, stride, dropRate=0.0):
super(NetworkBlock, 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, num_classes, widen_factor=1, 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 = BasicBlock
# 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 = NetworkBlock(n, nChannels[0], nChannels[1], block, 1, dropRate)
# 2nd block
self.block2 = NetworkBlock(n, nChannels[1], nChannels[2], block, 2, dropRate)
# 3rd block
self.block3 = NetworkBlock(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, math.sqrt(2. / n))
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
elif isinstance(m, nn.Linear):
m.bias.data.zero_()
def forward(self, x):
out = self.conv1(x)
out = self.block1(out)
activation1 = out
out = self.block2(out)
activation2 = out
out = self.block3(out)
activation3 = out
out = self.relu(self.bn1(out))
out = F.avg_pool2d(out, 8)
out = out.view(-1, self.nChannels)
return self.fc(out), activation1, activation2, activation3
def main(args):
transform = transforms.Compose([transforms.ColorJitter(),
transforms.RandomResizedCrop(32, (0.5, 1.5)),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.5071,0.4867, 0.4408), (0.2675,0.2565,0.2761))])
train_dataset = datasets.CIFAR100('./Datasets', train=True, transform=transform, download=True)
test_transform = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.5071,0.4867, 0.4408), (0.2675,0.2565,0.2761))])
test_dataset = datasets.CIFAR100('./Datasets', train=False, transform=test_transform, download=True)
train_loader = DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers)
test_loader = DataLoader(test_dataset, batch_size=args.batch_size, shuffle=False, num_workers=args.num_workers)
net_depth, net_width = int(args.select_model[4:6]), int(args.select_model[-1])
Teacher_model = WideResNet(depth=net_depth, num_classes=100, widen_factor=net_width, dropRate=0.3)
# Teacher_model = WideResNet(depth=40, num_classes=10, widen_factor=2, dropRate=0.0)
Teacher_model.cuda()
Teacher_model = torch.nn.DataParallel(Teacher_model, device_ids=range(torch.cuda.device_count()))
cudnn.benchmark = True
optimizer = torch.optim.SGD(Teacher_model.parameters(), lr=args.lr, momentum=0.9, weight_decay=5e-4, nesterov=True)
optim_scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, milestones=[60, 120, 160], gamma=0.2)
criterion = nn.CrossEntropyLoss()
best_acc = 0
for epoch in range(args.total_epochs):
for iter_, data in enumerate(train_loader):
images, labels = data
images, labels = images.to('cuda:0'), labels.to('cuda:0')
outputs, *acts = Teacher_model(images)
classification_loss = criterion(outputs, labels)
optimizer.zero_grad()
classification_loss.backward()
optimizer.step()
if iter_%10==0:
print("Epoch: {}/{}, Iteration: {}/{}, Loss: {:02.5f}".format(epoch,args.total_epochs, iter_, train_loader.__len__(), classification_loss.item()))
with torch.no_grad():
Teacher_model.eval()
cumulated_acc=0
for x, y in test_loader:
x, y = x.to('cuda:0'), y.to('cuda:0')
logits, *activations = Teacher_model(x)
acc = accuracy(logits.data, y, topk=(1,))[0]
cumulated_acc+=acc
print("Test Accuracy is {:02.2f} %".format(cumulated_acc/test_loader.__len__()*100))
Teacher_model.train()
if best_acc <= cumulated_acc/test_loader.__len__()*100:
best_acc = cumulated_acc/test_loader.__len__()*100
torch.save(Teacher_model.state_dict(),'./Pretrained/CIFAR100/WRN-{}-{}/Teacher_best.ckpt'.format(net_depth, net_width))
optim_scheduler.step()
if __name__=="__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--weight_path', type=str, default='')
parser.add_argument('--total_epochs', type=int, default=200)
parser.add_argument('--lr', type=float, default=1e-1)
parser.add_argument('--batch_size', type=int, default=128, help='total training batch size')
parser.add_argument('--select_model', type=str, default='WRN-40-2', help='What do you want to train?')
parser.add_argument('--num_workers', type=int, default=4)
args = parser.parse_args()
main(args)