-
Notifications
You must be signed in to change notification settings - Fork 0
/
aux_loss_main.py
273 lines (215 loc) · 10.5 KB
/
aux_loss_main.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
#!/usr/bin/env python
# coding: utf-8
# In[ ]:
#!/usr/bin/env python
# coding: utf-8
# !/usr/bin/python
# -*- coding: utf-8 -*-
import os
import time
import random
import argparse
import torch
import numpy as np
import timeit
import configparser
import torch.optim as optim
from tools.trainer_aux import Trainer
from tools.tester_aux import Tester
from torch.utils.data import DataLoader
from metrics.metrics_uavid import runningScore, averageMeter
import torch.backends.cudnn as cudnn
from utils.modeltools import netParams
from utils.set_logger import get_logger
import utils.utils
from network import build_network
import warnings
warnings.filterwarnings('ignore')
cfg = configparser.RawConfigParser()
cfg.read("config.ini")
dataset_type = cfg.get("init", "DATASET")
MODEL_INIT = cfg.get(dataset_type, "MODEL_INIT")
ROOT = cfg.get(dataset_type, "ROOT")
BATCH_SIZE = cfg.getint(dataset_type, "BATCH_SIZE")
MAX_EPOCHES = cfg.getint(dataset_type, "MAX_EPOCHES")
LR_INIT = cfg.getfloat(dataset_type, "LR_INIT")
NUM_CLASSES = cfg.getint(dataset_type, "NUM_CLASSES")
SAVE_DIR = cfg.get(dataset_type, "SAVE_DIR")
GPU = cfg.get(dataset_type, "GPU")
REPEAT_TIME = cfg.getint(dataset_type, "REPEAT")
RUN_ID = MODEL_INIT+'_'+str(MAX_EPOCHES)+'_'+str(REPEAT_TIME)
def main(args, logger):
cudnn.enabled = True # Enables bencnmark mode in cudnn, to enable the inbuilt
cudnn.benchmark = True # cudnn auto-tuner to find the best algorithm to use for
# our hardware
#Setup random seed
# cudnn.deterministic = True # ensure consistent results
# if benchmark = True, deterministic will be False.
seed = random.randint(1, 10000)
print('======>random seed {}'.format(seed))
logger.info('======>random seed {}'.format(seed))
random.seed(seed) # python random seed
np.random.seed(seed) # set numpy random seed
start = timeit.default_timer()
torch.manual_seed(seed) # set random seed for cpu
if torch.cuda.is_available():
# torch.cuda.manual_seed(seed) # set random seed for GPU now
torch.cuda.manual_seed_all(seed) # set random seed for all GPU
# Setup device
# device = torch.device("cuda:{}".format(args.gpu) if torch.cuda.is_available() else "cpu")
# setup DatasetLoader
if dataset_type == 'uavid':
from loader.load_uavid import uavidloader
train_set = uavidloader(root=args.root, split='train')
test_set = uavidloader(root=args.root, split='val')
elif dataset_type == 'udd6':
from loader.load_udd6 import udd6loader
train_set = udd6loader(root=args.root, split='train')
test_set = udd6loader(root=args.root, split='val')
elif dataset_type == 'vai':
from loader.load_vaihingen import vaihingenloader
train_set = vaihingenloader(root=args.root, split='train')
test_set = vaihingenloader(root=args.root, split='test')
else:
from loader.load_udd6 import udd6loader
train_set = udd6loader(root=args.root, split='train')
test_set = udd6loader(root=args.root, split='val')
kwargs = {'num_workers': args.workers, 'pin_memory': True}
train_loader = DataLoader(train_set, batch_size=args.batch_size, shuffle=True, **kwargs)
# test_loader = DataLoader(test_set, batch_size=args.batch_size, shuffle=False, **kwargs)
test_loader = DataLoader(test_set, batch_size=1, shuffle=False, **kwargs)
# setup optimization criterion
criterion = utils.utils.cross_entropy2d
# setup model
print('======> building network')
logger.info('======> building network')
model = build_network(MODEL_INIT, NUM_CLASSES)
if torch.cuda.device_count() > 1:
device_ids = list(map(int, args.gpu.split(',')))
model = torch.nn.DataParallel(model, device_ids=device_ids)
print("======> computing network parameters")
logger.info("======> computing network parameters")
total_paramters = netParams(model)
print("the number of parameters: " + str(total_paramters))
logger.info("the number of parameters: " + str(total_paramters))
# setup optimizer
optimizer = optim.SGD(model.parameters(), lr=args.lr, momentum=0.9, weight_decay=args.weight_decay)
# optimizer = torch.optim.Adam(model.parameters(), args.lr, (0.9, 0.999), eps=1e-08, weight_decay=5e-4)
# setup savedir
args.savedir = (args.savedir + '/' + args.model + 'bs'
+ str(args.batch_size) + 'gpu' + str(args.gpu) + '/')
if not os.path.exists(args.savedir):
os.makedirs(args.savedir)
start_epoch = 0
flag = True
best_epoch = 0.
best_overall = 0.
best_mIoU = 0.
best_F1 = 0.
epoch = 0
train_1_epoch = Trainer(args, train_loader, model, criterion, optimizer, epoch, logger)
testing = Tester(args, test_loader, model, criterion, optimizer, epoch, logger)
while flag == True:
for epoch in range(start_epoch, args.max_epochs):
print('======> Epoch {} starting train.'.format(epoch))
logger.info('======> Epoch {} starting train.'.format(epoch))
# print(logger)
epoch_start = timeit.default_timer()
train_1_epoch.train_net(epoch)
train_end = timeit.default_timer()
print("training time:", 1.0*(train_end-epoch_start)/3600)
print('======> Epoch {} train finish.'.format(epoch))
logger.info('======> Epoch {} train finish.'.format(epoch))
if epoch % 1 == 0 or (epoch + 1) == args.max_epochs:
print('Now Epoch {}, starting evaluate on Test dataset.'.format(epoch))
logger.info('Now starting evaluate on Test dataset.')
print('length of test set:', len(test_loader))
logger.info('length of test set: {}'.format(len(test_loader)))
score, class_iou, class_F1 = testing.test_net()
for k, v in score.items():
print('{}: {:.5f}'.format(k, v))
logger.info('======>{0:^18} {1:^10}'.format(k, v))
print('Now print class iou')
for k, v in class_iou.items():
print('{}: {:.5f}'.format(k, v))
logger.info('======>{0:^18} {1:^10}'.format(k, v))
print('Now print class_F1')
for k, v in class_F1.items():
print('{}: {:.5f}'.format(k, v))
logger.info('======>{0:^18} {1:^10}'.format(k, v))
if score["Mean IoU : \t"] > best_mIoU:
best_mIoU = score["Mean IoU : \t"]
model_file_name = args.savedir + '/model.pth'
torch.save(model.module.state_dict(), model_file_name)
best_epoch = epoch
if score["Overall Acc : \t"] > best_overall:
best_overall = score["Overall Acc : \t"]
# save model in best overall Acc
# TODO functionalize it
if score["Mean F1 : \t"] > best_F1:
best_F1 = score["Mean F1 : \t"]
print(f"best mean IoU: {best_mIoU}")
print(f"best overall : {best_overall}")
print(f"best F1: {best_F1}")
print(f"best epoch: {best_epoch}")
logger.info('best mean IoU: {}'.format(best_mIoU))
logger.info('best overall: {}'.format(best_overall))
logger.info('best F1: {}'.format(best_F1))
logger.info('best epoch: {}'.format(best_epoch))
epoch_end = timeit.default_timer()
print("evaluation time:", -1.0*(train_end - epoch_end)/3600)
# #save the model
# model_file_name = args.savedir +'/model.pth'
# state = {"epoch": epoch+1, "model": model.state_dict()}
#
# if (epoch + 1) == args.max_epochs or epoch % 5 == 0:
# print('======> Now begining to save model.')
# logger.info('======> Now begining to save model.')
# torch.save(state, model_file_name)
# print('======> Save done.')
# logger.info('======> Save done.')
#
if (epoch + 1) == args.max_epochs:
# print('the best pred mIoU: {}'.format(best_pred))
flag = False
break
if __name__ == '__main__':
import timeit
start = timeit.default_timer()
parser = argparse.ArgumentParser(description='Semantic Segmentation...')
parser.add_argument('--model', default=MODEL_INIT, type=str)
parser.add_argument('--root', default=ROOT, help='data directory')
parser.add_argument('--batch_size', default=BATCH_SIZE, type=int)
parser.add_argument('--max_epochs', type=int, default=MAX_EPOCHES, help='the number of epochs: default 100 ')
parser.add_argument('--num_classes', default=NUM_CLASSES, type=int)
parser.add_argument('--lr', default=LR_INIT, type=float)
parser.add_argument('--weight_decay', default=4e-5, type=float)
parser.add_argument('--workers', type=int, default=2, help=" the number of parallel threads")
parser.add_argument('--show_interval', default=10, type=int)
parser.add_argument('--show_val_interval', default=1000, type=int)
parser.add_argument('--savedir', default=SAVE_DIR, help="directory to save the model snapshot")
# parser.add_argument('--logFile', default= "log.txt", help = "storing the training and validation logs")
parser.add_argument('--gpu', type=str, default=GPU, help="default GPU devices (3)")
args = parser.parse_args()
print('Now run_id {}'.format(RUN_ID))
args.savedir = os.path.join(args.savedir, str(RUN_ID))
print(args.savedir)
if not os.path.exists(args.savedir):
os.makedirs(args.savedir)
logger = get_logger(args.savedir)
logger.info('just do it')
print('Input arguments:')
logger.info('======>Input arguments:')
for key, val in vars(args).items():
print('======>{:16} {}'.format(key, val))
logger.info('======> {:16} {}'.format(key, val))
if torch.cuda.device_count() > 1:
torch.cuda.set_device(int(args.gpu.split(',')[0]))
# os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu.split(',')[0]
else:
# os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu
torch.cuda.set_device(int(args.gpu))
main(args, logger)
end = timeit.default_timer()
print("training time:", 1.0*(end-start)/3600)
print('model save in {}.'.format(RUN_ID))