-
Notifications
You must be signed in to change notification settings - Fork 80
/
train.py
299 lines (255 loc) · 10.3 KB
/
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
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
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
# Code for KiU-Net
# Author: Jeya Maria Jose
import argparse
import torch
import torchvision
from torch import nn
from torch.autograd import Variable
from torch.utils.data import DataLoader
from torchvision import transforms
from torchvision.utils import save_image
import torch.nn.functional as F
import os
import matplotlib.pyplot as plt
import torch.utils.data as data
from PIL import Image
import numpy as np
from torchvision.utils import save_image
import torch
import torch.nn.init as init
from utils import JointTransform2D, ImageToImage2D, Image2D
from metrics import jaccard_index, f1_score, LogNLLLoss,classwise_f1
from utils import chk_mkdir, Logger, MetricList
import cv2
from functools import partial
from random import randint
import timeit
from arch.ae import kiunet,kinetwithsk,unet,autoencoder, reskiunet,densekiunet, kiunet3d
def mae(imageA, imageB):
# the 'Mean Squared Error' between the two images is the
# sum of the squared difference between the two images;
# NOTE: the two images must have the same dimension
err = np.sum(abs(imageA.astype("float") - imageB.astype("float")) )
err /= float(imageA.shape[0] * imageA.shape[1])
# return the MSE, the lower the error, the more "similar"
# the two images are
return err
parser = argparse.ArgumentParser(description='KiU-Net')
parser.add_argument('-j', '--workers', default=16, type=int, metavar='N',
help='number of data loading workers (default: 8)')
parser.add_argument('--epochs', default=100, type=int, metavar='N',
help='number of total epochs to run(default: 1)')
parser.add_argument('--start-epoch', default=0, type=int, metavar='N',
help='manual epoch number (useful on restarts)')
parser.add_argument('-b', '--batch_size', default=1, type=int,
metavar='N', help='batch size (default: 8)')
parser.add_argument('--learning_rate', default=1e-3, type=float,
metavar='LR', help='initial learning rate (default: 0.01)')
parser.add_argument('--momentum', default=0.9, type=float, metavar='M',
help='momentum')
parser.add_argument('--weight-decay', '--wd', default=1e-5, type=float,
metavar='W', help='weight decay (default: 1e-4)')
parser.add_argument('--lfw_path', default='../lfw', type=str, metavar='PATH',
help='path to root path of lfw dataset (default: ../lfw)')
parser.add_argument('--train_dataset', required=True, type=str)
parser.add_argument('--val_dataset', type=str)
parser.add_argument('--save_freq', type=int,default = 5)
parser.add_argument('--modelname', default='off', type=str,
help='turn on img augmentation (default: False)')
parser.add_argument('--cuda', default="on", type=str,
help='switch on/off cuda option (default: off)')
parser.add_argument('--aug', default='off', type=str,
help='turn on img augmentation (default: False)')
parser.add_argument('--load', default='default', type=str,
help='turn on img augmentation (default: default)')
parser.add_argument('--save', default='default', type=str,
help='turn on img augmentation (default: default)')
parser.add_argument('--model', default='kiunet', type=str,
help='model name')
parser.add_argument('--direc', default='./brainus_OC_udenet', type=str,
help='directory to save')
parser.add_argument('--crop', type=int, default=None)
parser.add_argument('--device', default='cuda', type=str)
parser.add_argument('--edgeloss', default='off', type=str)
args = parser.parse_args()
aug = args.aug
direc = args.direc
modelname = args.modelname
losstype= args.edgeloss
def weight_init(m):
'''
Usage:
model = Model()
model.apply(weight_init)
'''
if isinstance(m, nn.Conv1d):
init.normal_(m.weight.data)
if m.bias is not None:
init.normal_(m.bias.data)
elif isinstance(m, nn.Conv2d):
init.xavier_normal_(m.weight.data)
if m.bias is not None:
init.normal_(m.bias.data)
elif isinstance(m, nn.Conv3d):
init.xavier_normal_(m.weight.data)
if m.bias is not None:
init.normal_(m.bias.data)
elif isinstance(m, nn.ConvTranspose1d):
init.normal_(m.weight.data)
if m.bias is not None:
init.normal_(m.bias.data)
elif isinstance(m, nn.ConvTranspose2d):
init.xavier_normal_(m.weight.data)
if m.bias is not None:
init.normal_(m.bias.data)
elif isinstance(m, nn.ConvTranspose3d):
init.xavier_normal_(m.weight.data)
if m.bias is not None:
init.normal_(m.bias.data)
def add_noise(img):
#random noise
noise = torch.randn(img.size()) * 0.1
noisy_img = img + noise.cuda()
return noisy_img
if args.crop is not None:
crop = (args.crop, args.crop)
else:
crop = None
tf_train = JointTransform2D(crop=crop, p_flip=0.5, color_jitter_params=None, long_mask=True)
tf_val = JointTransform2D(crop=crop, p_flip=0, color_jitter_params=None, long_mask=True)
train_dataset = ImageToImage2D(args.train_dataset, tf_val)
val_dataset = ImageToImage2D(args.val_dataset, tf_val)
predict_dataset = Image2D(args.val_dataset)
dataloader = DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True)
valloader = DataLoader(val_dataset, 1, shuffle=True)
device = torch.device("cuda")
if modelname == "unet":
model = unet()
elif modelname =="autoencoder":
model =autoencoder()
elif modelname == "kiunet":
model = kiunet()
elif modelname == "kinetwithsk":
model = kinetwithsk()
elif modelname == "kinet":
model = kinet()
elif modelname == "reskiunet":
model = reskiunet()
elif modelname == "densekiunet":
model = densekiunet()
elif modelname == "kiunet3d":
model = kiunder3d()
elif modelname == "pspnet":
model = psp.PSPNet(layers=5, bins=(1, 2, 3, 6), dropout=0.1, classes=21, zoom_factor=1, use_ppm=True, pretrained=False).cuda()
if torch.cuda.device_count() > 1:
print("Let's use", torch.cuda.device_count(), "GPUs!")
# dim = 0 [30, xxx] -> [10, ...], [10, ...], [10, ...] on 3 GPUs
model = nn.DataParallel(model,device_ids=[0,1]).cuda()
model.to(device)
# model.apply(weight_init)
# print(model)
bestdice=0
criterion = LogNLLLoss()
optimizer = torch.optim.Adam(list(model.parameters()), lr=args.learning_rate,
weight_decay=1e-5)
metric_list = MetricList({'jaccard': partial(jaccard_index),
'f1': partial(f1_score)})
pytorch_total_params = sum(p.numel() for p in model.parameters() if p.requires_grad)
print("Total_params: {}".format(pytorch_total_params))
for epoch in range(args.epochs):
# break
epoch_running_loss = 0
for batch_idx, (X_batch, y_batch, *rest) in enumerate(dataloader):
###augmentations
X_batch = Variable(X_batch.to(device ='cuda'))
y_batch = Variable(y_batch.to(device='cuda'))
numr = randint(0,9)
if aug=='on':
if numr == 2:
# print(X_batch,y_batch)
X_batch = torch.flip(X_batch,[2,3])
y_batch = torch.flip(y_batch,[1,2])
# print(X_batch,y_batch)
elif numr ==3:
X_batch = torch.flip(X_batch,[3,2])
y_batch = torch.flip(y_batch,[2,1])
elif numr==4:
X_batch = add_noise(X_batch)
# y_batch = add_noise(y_batch)
# noisy_in = add_noise(X_batch)
# ===================forward=====================
output = model(X_batch)
tmp2 = y_batch.detach().cpu().numpy()
tmp = output.detach().cpu().numpy()
tmp[tmp>=0.5] = 1
tmp[tmp<0.5] = 0
tmp2[tmp2>0] = 1
tmp2[tmp2<=0] = 0
tmp2 = tmp2.astype(int)
tmp = tmp.astype(int)
# print(np.unique(tmp2))
yHaT = tmp
yval = tmp2
if losstype is 'on':
edgeloss = mae(yHaT,yval)
else:
edgeloss = 0
loss = criterion(output, y_batch)
loss =loss + edgeloss/10000
# ===================backward====================
optimizer.zero_grad()
loss.backward()
optimizer.step()
epoch_running_loss += loss.item()
# break
# ===================log========================
print('epoch [{}/{}], loss:{:.4f}'
.format(epoch, args.epochs, epoch_running_loss/(batch_idx+1)))
# =================validation=============
# metric_list.reset()
tf1 = 0
tmiou = 0
tpa = 0
count = 0
if (epoch % args.save_freq) ==0:
for batch_idx, (X_batch, y_batch, *rest) in enumerate(valloader):
# print(batch_idx)
if isinstance(rest[0][0], str):
image_filename = rest[0][0]
else:
image_filename = '%s.png' % str(batch_idx + 1).zfill(3)
X_batch = Variable(X_batch.to(device='cuda'))
y_batch = Variable(y_batch.to(device='cuda'))
# start = timeit.default_timer()
y_out = model(X_batch)
# stop = timeit.default_timer()
# print('Time: ', stop - start)
tmp2 = y_batch.detach().cpu().numpy()
tmp = y_out.detach().cpu().numpy()
tmp[tmp>=0.5] = 1
tmp[tmp<0.5] = 0
tmp2[tmp2>0] = 1
tmp2[tmp2<=0] = 0
tmp2 = tmp2.astype(int)
tmp = tmp.astype(int)
# print(np.unique(tmp2))
yHaT = tmp
yval = tmp2
epsilon = 1e-20
del X_batch, y_batch,tmp,tmp2, y_out
count = count + 1
yHaT[yHaT==1] =255
yval[yval==1] =255
fulldir = direc+"/{}/".format(epoch)
# print(fulldir+image_filename)
if not os.path.isdir(fulldir):
os.makedirs(fulldir)
cv2.imwrite(fulldir+image_filename, yHaT[0,1,:,:])
# cv2.imwrite(fulldir+'/gt_{}.png'.format(count), yval[0,:,:])
fulldir = direc+"/{}/".format(epoch)
torch.save(model.state_dict(), fulldir+args.model+".pth")
torch.save(model.state_dict(), direc+"model.pth")
if bestdice<tf1:
bestdice = tf1
print("bestdice = {}".format(bestdice/count))
print(epoch)