-
Notifications
You must be signed in to change notification settings - Fork 10
/
Copy pathresnet50nodown.py
204 lines (162 loc) · 7.16 KB
/
resnet50nodown.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
# -*- coding: utf-8 -*-
# %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
#
# Copyright (c) 2019 Image Processing Research Group of University Federico II of Naples ('GRIP-UNINA').
# All rights reserved.
# This work should only be used for nonprofit purposes.
#
# By downloading and/or using any of these files, you implicitly agree to all the
# terms of the license, as specified in the document LICENSE.md
# (included in this package) and online at
# http://www.grip.unina.it/download/LICENSE_OPEN.txt
#
import torch
import torch.nn as nn
import torchvision.transforms as transforms
LIMIT_SIZE = 1536
LIMIT_SLIDE = 1024
class ChannelLinear(nn.Linear):
def __init__(self, in_features: int, out_features: int, bias: bool = True) -> None:
super(ChannelLinear, self).__init__(in_features, out_features, bias)
def forward(self, x):
out_shape = [x.shape[0], x.shape[2], x.shape[3], self.out_features]
x = x.permute(0,2,3,1).reshape(-1,self.in_features)
x = x.matmul(self.weight.t())
if self.bias is not None:
x = x + self.bias[None,:]
x = x.view(out_shape).permute(0,3,1,2)
return x
def conv3x3(in_planes, out_planes, stride=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False)
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 Bottleneck(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(Bottleneck, self).__init__()
self.conv1 = conv1x1(inplanes, planes)
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = conv3x3(planes, planes, stride)
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = conv1x1(planes, planes * self.expansion)
self.bn3 = nn.BatchNorm2d(planes * self.expansion)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
def forward(self, x):
identity = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None:
identity = self.downsample(x)
out += identity
out = self.relu(out)
return out
class ResNet(nn.Module):
def __init__(self, block, layers, num_classes=1, stride0=2):
super(ResNet, self).__init__()
self.inplanes = 64
self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=stride0, padding=3, bias=False)
self.bn1 = nn.BatchNorm2d(64)
self.relu = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=stride0, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.num_features = 512 * block.expansion
self.fc = ChannelLinear(self.num_features, num_classes)
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
elif isinstance(m, nn.BatchNorm2d):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
# transform form Pillow
self.transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])])
def _make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
conv1x1(self.inplanes, planes * block.expansion, stride),
nn.BatchNorm2d(planes * block.expansion),
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for _ in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
def change_output(self, num_classes):
self.fc = ChannelLinear(self.num_features, num_classes)
torch.nn.init.normal_(self.fc.weight.data, 0.0, 0.02)
return self
def change_input(self, num_inputs):
data = self.conv1.weight.data
old_num_inputs = int(data.shape[1])
if num_inputs>old_num_inputs:
times = num_inputs//old_num_inputs
if (times*old_num_inputs)<num_inputs:
times = times+1
data = data.repeat(1,times,1,1) / times
elif num_inputs==old_num_inputs:
return self
data = data[:,:num_inputs,:,:]
print(self.conv1.weight.data.shape, '->', data.shape)
self.conv1.weight.data = data
return self
def feature(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.maxpool(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.layer4(x)
return x
def forward(self, x):
x = self.feature(x)
x = self.avgpool(x)
x = self.fc(x)
return x
def apply(self, pil):
device = self.conv1.weight.device
if (pil.size[0]>LIMIT_SIZE) and (pil.size[1]>LIMIT_SIZE):
import numpy as np
print('err:', pil.size)
with torch.no_grad():
img = self.transform(pil)
list_logit = list()
list_weight = list()
for index0 in range(0, img.shape[-2], LIMIT_SLIDE):
for index1 in range(0, img.shape[-1], LIMIT_SLIDE):
clip = img[..., index0:min(index0+LIMIT_SLIDE, img.shape[-2]),
index1:min(index1+LIMIT_SLIDE, img.shape[-1])]
logit = torch.squeeze(self(clip.to(device)[None,:,:,:])).cpu().numpy()
weight = clip.shape[-2] * clip.shape[-1]
list_logit.append(logit)
list_weight.append(weight)
logit = np.mean(np.asarray(list_logit) * np.asarray(list_weight)) / np.mean(list_weight)
else:
with torch.no_grad():
logit = torch.squeeze(self(self.transform(pil).to(device)[None,:,:,:])).cpu().numpy()
return logit
def resnet50nodown(device, filename, num_classes=1):
"""Constructs a ResNet-50 nodown model.
"""
model = ResNet(Bottleneck, [3, 4, 6, 3], num_classes=num_classes, stride0=1)
model.load_state_dict(torch.load(filename, map_location=torch.device('cpu'))['model'])
model = model.to(device).eval()
return model