-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathresnet_attention.py
205 lines (176 loc) · 6.48 KB
/
resnet_attention.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
205
import torch
import torch.nn as nn
from torch.nn import init
from torchvision import models
from torch.autograd import Variable
import torch.nn.functional as F
from torchvision.utils import save_image
from torchvision import transforms
######################################################################
def weights_init_kaiming(m):
classname = m.__class__.__name__
# print(classname)
if classname.find('Conv') != -1:
init.kaiming_normal_(m.weight.data, a=0, mode='fan_in')
elif classname.find('Linear') != -1:
init.kaiming_normal_(m.weight.data, a=0, mode='fan_out')
init.constant_(m.bias.data, 0.0)
elif classname.find('BatchNorm1d') != -1:
init.normal_(m.weight.data, 1.0, 0.02)
init.constant_(m.bias.data, 0.0)
def weights_init_classifier(m):
classname = m.__class__.__name__
if classname.find('Linear') != -1:
init.normal_(m.weight.data, std=0.001)
init.constant_(m.bias.data, 0.0)
# Defines the new fc layer and classification layer
# |--Linear--|--bn--|--relu--|--Linear--|
class ClassBlock(nn.Module):
def __init__(self, input_dim, class_num, dropout=True, relu=True, num_bottleneck=512):
super(ClassBlock, self).__init__()
add_block = []
add_block += [nn.Linear(input_dim, num_bottleneck)]
add_block += [nn.BatchNorm1d(num_bottleneck)]
if relu:
add_block += [nn.LeakyReLU(0.1)]
if dropout:
add_block += [nn.Dropout(p=0.5)]
add_block = nn.Sequential(*add_block)
add_block.apply(weights_init_kaiming)
classifier = []
classifier += [nn.Linear(num_bottleneck, class_num)]
classifier = nn.Sequential(*classifier)
classifier.apply(weights_init_classifier)
self.add_block = add_block
self.classifier = classifier
def forward(self, x):
x = self.add_block(x)
x = self.classifier(x)
return x
# Define the ResNet50-based Model#
class ResNetAttention(nn.Module):
def __init__(self, num_class):
super(ResNetAttention, self).__init__()
model_ft = models.resnet50(pretrained=True)
# avg pooling to global pooling
model_ft.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.model = model_ft
self.classifier = ClassBlock(2048, num_class)
self.conv_attention = nn.Sequential(
nn.Conv2d(2048, 1, 1),
nn.BatchNorm2d(1),
nn.ReLU(inplace=True)
)
#nn.init.kaiming_normal_(self.conv_attention.weight)
self.conv_attention.apply(weights_init_kaiming)
#self.conv_proc_detail = nn.Sequential(
# nn.Conv2d(2048, 2048, 3, stride=1, padding=1),
# nn.BatchNorm2d(2048),
# nn.ReLU(inplace=True)
#)
#nn.init.kaiming_normal_(self.conv_proc_detail.weight)
#self.conv_proc_detail.apply(weights_init_kaiming)
def reprocess(self, x):
x = self.model.conv1(x)
x = self.model.bn1(x)
x = self.model.relu(x)
x = self.model.maxpool(x)
x = self.model.layer1(x)
x = self.model.layer2(x)
x = self.model.layer3(x)
x = self.model.layer4(x)
self.pool4 = F.max_pool2d(x, 2, 2)
def attend(self, x):
x = self.model.conv1(x)
x = self.model.bn1(x)
x = self.model.relu(x)
x = self.model.maxpool(x)
x = self.model.layer1(x)
x = self.model.layer2(x)
x = self.model.layer3(x)
x = self.model.layer4(x)
self.pool = F.max_pool2d(x, 2, 2)
b, c, h, w = self.pool.size()
b, c, h, w = x.size()
att = self.conv_attention(x)
#att = F.upsample(att.view(b, 1, h, w), scale_factor=2, mode='bilinear', align_corners=True)
att_mask = F.softmax(att.view(b, 1 * h * w), -1).view(b, 1, h, w)
return att_mask
def crop(self, x, h_multiple=4, w_multiple=2):
""" Adjusts the high resolution image feature size to be multiple of 7
Args:
x: input features
multiple: multiple to adjust to
Returns: cropped feature map
"""
b, c, h, w = x.size()
h_ = h % h_multiple
w_ = w % w_multiple
return x[:, :, 0:(h - h_), 0:(w - w_)]
def forward(self, x):
b1, c1, h1, w1 = x.size()
x2 = F.upsample(x, scale_factor=2)
b2, c2, h2, w2 = x2.size()
att = self.attend(x)
if (h1, w1) != (h2, w2):
self.reprocess(x2)
x_high = self.pool4
b, c, h, w = x_high.size()
b_att, c_att, h_att, w_att = att.size()
#x_high = F.max_pool2d(x_high, 2, 2)
#x_high = x_high[:, :, 0:(h - h_att), 0:(w - w_att)]
#print(x_high.size())
#print(att.size())
#exit(0)
# x_high = self.conv_proc_detail(x_high)
#att = F.upsample(att, scale_factor=2, mode='bilinear', align_corners=True)
attended = x_high * att
#attended = F.normalize(attended.view(b, c, -1).sum(-1), 2, -1)
x = self.model.avgpool(attended)
x = F.normalize(x.view(b, c, -1).sum(-1), 2, -1)
#x = torch.squeeze(x)
x = self.classifier(x)
return x
'''
x = att.view(b, 1, h, w)
x = x.detach().cpu()
#x = x.detach().cpu().numpy()
import torch
#x = torch.randn(32, 1, 3, 3)#
#print(type(x))
#print(x.size())
#from torchvision import transforms
#transform = transforms.Compose([
# transforms.ToPILImage(),
# transforms.Resize(size=24),
# transforms.ToTensor()
#])
#x = [transform(x_) for x_ in x]
#save_image(x, './images/att.png')
print(x.size())
img = x.detach().cpu().numpy()
img = img[0]
img = img[0]
import numpy as np
img = np.transpose(img, (1, 0))
print(img.shape)
from scipy.misc import imsave
imsave("./images/1dimconv1.png", img)
exit(0)
save_image(x, "images/conv1.png")
'''
# Define the DenseNet121-based Model
class DenseNetModel(nn.Module):
def __init__(self, class_num ):
super().__init__()
model_ft = models.densenet121(pretrained=True)
model_ft.features.avgpool = nn.AdaptiveAvgPool2d((1,1))
model_ft.fc = nn.Sequential()
self.model = model_ft
# For DenseNet, the feature dim is 1024
self.classifier = ClassBlock(1024, class_num)
def forward(self, x):
x = self.model.features(x)
x = torch.squeeze(x)
x = self.classifier(x)
return x