/
main.py
68 lines (55 loc) · 2.44 KB
/
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
import os
import torch
import argparse
from torch.backends import cudnn
from models.Net import build_net
from train import _train
from test import _test
import torch.nn as nn
def main(args):
# CUDNN
cudnn.benchmark = True
if not os.path.exists(args.model_save_dir):
os.makedirs(args.model_save_dir)
model = build_net(args.model_name)
if torch.cuda.is_available():
gpus = [i for i in range(args.gpu_device)]
device = torch.device('cuda:{}'.format(gpus[0]))
torch.cuda.set_device(device)
model = model.cuda()
model = nn.DataParallel(model.to(device), device_ids=gpus, output_device=gpus[0])
if args.mode == 'train':
_train(model, args)
elif args.mode == 'test':
_test(model, args)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
# Directories
parser.add_argument('--model_name', default='RFLLIE', type=str)
parser.add_argument('--data_dir', type=str,
default='dataset/LOLv1/', help='training data')
parser.add_argument('--test_dir', type=str,
default='dataset/LOLv1/val', help='evaluation data')
parser.add_argument('--mode', default='train', choices=['train', 'test'], type=str)
# Train
parser.add_argument('--gpu_device', default=2, type=int,
help='4 for training and 1 for evaluation')
parser.add_argument('--batch_size', type=int, default=8)
parser.add_argument('--learning_rate', type=float, default=1e-4)
parser.add_argument('--weight_decay', type=float, default=0)
parser.add_argument('--num_epoch', type=int, default=200)
parser.add_argument('--print_freq', type=int, default=10)
parser.add_argument('--num_worker', type=int, default=4)
parser.add_argument('--save_freq', type=int, default=1)
parser.add_argument('--valid_freq', type=int, default=1)
parser.add_argument('--resume', type=str, default='')
parser.add_argument('--gamma', type=float, default=0.5)
parser.add_argument('--lr_steps', type=list, default=[(x+1) * 40 for x in range(200//40)])
parser.add_argument('--model_save_dir', default='ckpt/', type=str)
# Test
parser.add_argument('--test_model', type=str, default='ckpt/Best.pkl')
parser.add_argument('--save_image', type=bool, default=True, choices=[True, False])
parser.add_argument('--output_dir', type=str, default='results/')
args = parser.parse_args()
print(args)
main(args)