-
Notifications
You must be signed in to change notification settings - Fork 4
/
resnet_unet.py
197 lines (147 loc) · 6.92 KB
/
resnet_unet.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
import torch
import torch.nn as nn
from torchvision import models
import torch.nn.functional as F
import torch.optim as optim
from collections import defaultdict
import time
import copy
'''
Helpful Links:
Conv2D layers: https://pytorch.org/docs/stable/nn.html#conv2d
Upsampling layers: https://pytorch.org/docs/stable/_modules/torch/nn/modules/upsampling.html
'''
def convrelu(in_channels, out_channels, kernel_size, padding):
return nn.Sequential(
nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding),
nn.ReLU(inplace=True)
)
class ResNetUNet(nn.Module):
def __init__(self, num_classes):
super().__init__()
self.base_model = models.resnet18(pretrained=True)
self.base_layers = list(self.base_model.children())
self.layer0 = nn.Sequential(*self.base_layers[:3]) # size=(N, 64, x.H/2 x.W/2)TODO:?
self.layer0_1x1 = convrelu(64, 64, 1, 0)
self.layer1 = nn.Sequential(*self.base_layers[3:5]) # size=(N, 64, x.H/4, x.W/4)
self.layer1_1x1 = convrelu(64, 64, 1, 0)
self.layer2 = self.base_layers[5] # size=(N, 128, x.H/8, x.W/8)
self.layer2_1x1 = convrelu(128, 128, 1, 0)
self.layer3 = self.base_layers[6] # size=(N, 256, x.H/16, x.W/16)
self.layer3_1x1 = convrelu(256, 256, 1, 0)
self.layer4 = self.base_layers[7] # size=(N, 512, x.H/32, x.W/32)
self.layer4_1x1 = convrelu(512, 512, 1, 0)
self.upsample = nn.Upsample(scale_factor=2, mode='bilinear', align_corners=True)
self.conv_up3 = convrelu(256 + 512, 512, 3, 1)
self.conv_up2 = convrelu(128 + 512, 256, 3, 1)
self.conv_up1 = convrelu(64 + 256, 256, 3, 1)
self.conv_up0 = convrelu(64 + 256, 128, 3, 1)
self.conv_original_size0 = convrelu(3, 64, 3, 1)
self.conv_original_size1 = convrelu(64, 64, 3, 1)
self.conv_original_size2 = convrelu(64 + 128, 64, 3, 1)
self.conv_last = nn.Conv2d(64, num_classes, 1)
def forward(self, input):
x_original = self.conv_original_size0(input) # N, 64, H, W
x_original = self.conv_original_size1(x_original) # N, 64, H, W
layer0 = self.layer0(input) # N, 64, H/2, W/2
layer1 = self.layer1(layer0) # N, 64, H/4, W/4
layer2 = self.layer2(layer1) # N, 128, H/8, W/8
layer3 = self.layer3(layer2) # N, 256, H/16, W/16
layer4 = self.layer4(layer3) # N, 512, H/32, W/32
layer4 = self.layer4_1x1(layer4) # N, 512, H/32, W/32
x = self.upsample(layer4) # N, 512, H/16, W/16
layer3 = self.layer3_1x1(layer3) # COPY LAYER 3: N, 256, H/16, W/16
x = torch.cat([x, layer3], dim=1) # N, 512 + 256, H/16, W/16
x = self.conv_up3(x) # N, 512, H/16?, W/16?
x = self.upsample(x) # N, 512, H/8, W/8
layer2 = self.layer2_1x1(layer2) # COPY LAYER 2: N, 128, H/8, W/8
x = torch.cat([x, layer2], dim=1) # N, 512 + 128, H/8, W/8
x = self.conv_up2(x) # N, 256,, H/8, W/8
x = self.upsample(x) # N, 256, H/4, W/4
layer1 = self.layer1_1x1(layer1) # COPY LAYER 1: N, 64, H/4, W/4
x = torch.cat([x, layer1], dim=1) # N, 256 + 64, H/4, W/4
x = self.conv_up1(x) # N, 256, H/4, W/4
x = self.upsample(x) # N, 256, H/2, W/2
layer0 = self.layer0_1x1(layer0) # COPY LAYER 0: N, 64, H/2, W/2
x = torch.cat([x, layer0], dim=1) # N, 256 + 64, H/2, W/2
x = self.conv_up0(x) # N, 128, H/2, W/2
x = self.upsample(x) # N, 128, H, W
x = torch.cat([x, x_original], dim=1) # N, 128 + 64, H, W
x = self.conv_original_size2(x) # N, 64, H, W
out = self.conv_last(x) # N, num_classes, H, Ws
return out
def load(self, PATH='checkpoint.pth'):
try:
self.load_state_dict(torch.load(PATH))
except FileNotFoundError:
print('Error while trying to load model. No file found for', PATH)
def save(self, PATH='checkpoint.pth'):
torch.save(self.state_dict(), PATH)
def train_model(model, optimizer, scheduler, num_epochs, dataloaders, device):
best_model_wts = copy.deepcopy(model.state_dict())
best_loss = 1e10
for epoch in range(num_epochs):
print('Epoch {}/{}'.format(epoch, num_epochs - 1))
print('-' * 10)
since = time.time()
# Each epoch has a training and validation phase
for phase in ['train', 'val']:
if phase == 'train':
scheduler.step()
for param_group in optimizer.param_groups:
print("LR", param_group['lr'])
model.train() # Set model to training mode
else:
model.eval() # Set model to evaluate mode
metrics = defaultdict(float)
epoch_samples = 0
for inputs, labels in dataloaders[phase]:
inputs = inputs.to(device)
labels = labels.to(device)
# zero the parameter gradients
optimizer.zero_grad()
# forward
# track history if only in train
with torch.set_grad_enabled(phase == 'train'):
outputs = model(inputs)
loss = calc_loss(outputs, labels, metrics)
# backward + optimize only if in training phase
if phase == 'train':
loss.backward()
optimizer.step()
# statistics
epoch_samples += inputs.size(0)
print_metrics(metrics, epoch_samples, phase)
epoch_loss = metrics['loss'] / epoch_samples
# deep copy the model
if phase == 'val' and epoch_loss < best_loss:
print("saving best model")
best_loss = epoch_loss
best_model_wts = copy.deepcopy(model.state_dict())
torch.save(model.state_dict(), 'checkpoint.pth')
time_elapsed = time.time() - since
print('{:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
print('Best val loss: {:4f}'.format(best_loss))
# load best model weights
model.load_state_dict(best_model_wts)
return model
def dice_loss(pred, target, smooth = 1.):
pred = pred.contiguous()
target = target.contiguous()
intersection = (pred * target).sum(dim=2).sum(dim=2)
loss = (1 - ((2. * intersection + smooth) / (pred.sum(dim=2).sum(dim=2) + target.sum(dim=2).sum(dim=2) + smooth)))
return loss.mean()
def calc_loss(pred, target, metrics, bce_weight=0.5):
bce = F.binary_cross_entropy_with_logits(pred.double(), target.double())
pred = F.sigmoid(pred.double())
dice = dice_loss(pred.double(), target.double())
loss = bce * bce_weight + dice * (1 - bce_weight)
metrics['bce'] += bce.data.cpu().numpy() * target.size(0)
metrics['dice'] += dice.data.cpu().numpy() * target.size(0)
metrics['loss'] += loss.data.cpu().numpy() * target.size(0)
return loss
def print_metrics(metrics, epoch_samples, phase):
outputs = []
for k in metrics.keys():
outputs.append("{}: {:4f}".format(k, metrics[k] / epoch_samples))
print("{}: {}".format(phase, ", ".join(outputs)))