-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathocta_train.py
145 lines (113 loc) · 5.41 KB
/
octa_train.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
import argparse
import os.path
import cv2
import numpy
import torch
import torch.nn as nn
import torch.nn.functional as F
from vnet import VNetProjnl
import data, utils
def weight_mse(input, target, alpha=4.5, beta=0.5, gamma=2, epsilon=0.05, reduction='sum'):
mse = (input - target) ** 2
input_data = input.detach()
input_data[input_data < 0] = 0
input_data[input_data > 1] = 1
target_data = target.detach()
weights = alpha*torch.pow(input_data, 1.0/gamma) + beta*torch.pow(target_data, 1.0/gamma) + epsilon
weighted_mse = weights*mse
if reduction == 'sum':
return torch.sum(weighted_mse)
elif reduction == 'mean':
return torch.mean(weighted_mse)
else:
return weighted_mse
def main(args):
device = torch.device('cuda:0') if torch.cuda.is_available() else torch.device('cpu')
alpha = args.alpha
beta = args.beta
gamma = args.gamma
epsilon = args.epsilon
model = VNetProjnl(1, 1, args.n_frames)
model = model.to(device)
mid = args.n_frames // 2
optimizer = torch.optim.Adam(model.parameters(), lr=args.lr)
scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, milestones=[4, 8, 12], gamma=0.1)
global_step = -1
start_epoch = 0
train_loader, valid_loader = data.build_dataset(args.dataset, args.data_path, batch_size=args.batch_size,
image_size=args.image_size, stride=args.stride,
n_frames=args.n_frames, padding=args.padding)
# Track moving average of loss values
train_meters = {name: utils.RunningAverageMeter(0.98) for name in (["train_loss"])}
for epoch in range(start_epoch, args.num_epochs):
train_bar = utils.ProgressBar(train_loader, epoch)
for meter in train_meters.values():
meter.reset()
for batch_id, (inputs, targets, flags) in enumerate(train_bar):
model.train()
flags = flags.squeeze(1)
global_step += 1
inputs[:, :, mid, :, :] = 0
inputs = inputs.to(device)
targets = targets.to(device)
out = model(inputs)
if torch.sum(flags) != 0:
loss = weight_mse(out[flags], targets[flags], alpha=alpha, beta=beta, gamma=gamma, epsilon=epsilon, reduction='sum') / torch.sum(flags).item()
else:
model.zero_grad()
continue
model.zero_grad()
loss.backward()
optimizer.step()
if torch.sum(flags) != 0:
train_meters["train_loss"].update(loss.item())
train_bar.log(dict(**train_meters, lr=optimizer.param_groups[0]["lr"]), verbose=True)
torch.save(model.state_dict(), 'ckpt.pth')
scheduler.step()
if (epoch + 1) % args.valid_interval == 0:
model.eval()
if not os.path.exists('val_out'):
os.mkdir('val_out')
valid_bar = utils.ProgressBar(valid_loader)
for sample_id, (sample, target_name) in enumerate(valid_bar):
with torch.no_grad():
sample[:, :, mid, :, :] = 0
noisy_inputs = sample.to(device)
out = model(noisy_inputs)
img = out.cpu().squeeze().numpy()
img[img < 0] = 0
img[img > 1] = 1
img = img * 255
img = img.astype('uint8')
volume_id = target_name[0].split('_')[0] + '_val'
save_dir = os.path.join('val_out', volume_id)
if not os.path.exists(save_dir):
os.mkdir(save_dir)
cv2.imwrite(os.path.join(save_dir, target_name[0]), img)
if epoch + 1 == args.num_epochs:
torch.save(model.state_dict(), 'ckpt_' + str(epoch).zfill(2) + '.pth')
def get_args():
parser = argparse.ArgumentParser(allow_abbrev=False)
# Add data arguments
parser.add_argument("--data-path", default="data", help="path to data directory")
parser.add_argument("--dataset", default="OCTA", help="train dataset name")
parser.add_argument("--batch-size", default=128, type=int, help="train batch size")
parser.add_argument("--image-size", default=128, type=int, help="image size for train")
parser.add_argument("--n-frames", default=7, type=int, help="number of frames for training")
parser.add_argument("--stride", default=64, type=int, help="stride for patch extraction")
# Add optimization arguments
parser.add_argument("--lr", default=1e-3, type=float, help="learning rate")
parser.add_argument("--num-epochs", default=15, type=int, help="force stop training at specified epoch")
parser.add_argument("--valid-interval", default=1, type=int, help="evaluate every N epochs")
# Add loss arguments
parser.add_argument("--alpha", default=100, type=float, help="alpha")
parser.add_argument("--beta", default=1, type=float, help="beta")
parser.add_argument("--gamma", default=3, type=float, help="gamma")
parser.add_argument("--epsilon", default=0.5, type=float, help="epsilon")
# Add validation arguments
parser.add_argument("--padding", action='store_true', help="whether to replicate the boundary B-scans during validation")
args = parser.parse_args()
return args
if __name__ == "__main__":
args = get_args()
main(args)