-
Notifications
You must be signed in to change notification settings - Fork 27
/
Copy pathtrain.py
174 lines (149 loc) · 7.53 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
import argparse
import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, models, transforms
from dataloader import TrainDataset, collater, Resizer, PadToSquare,Color,Rotate,RandomErasing,RandomFlip, ValDataset
from torch.utils.data import Dataset, DataLoader, random_split
from terminaltables import AsciiTable, DoubleTable, SingleTable
# from tensorboardX import SummaryWriter
from torch.optim import lr_scheduler
import torch.distributed as dist
import eval_widerface
import torchvision
import model
import os
from torch.utils.data.distributed import DistributedSampler
import torchvision_model
def get_args():
parser = argparse.ArgumentParser(description="Train program for retinaface.")
parser.add_argument('--data_path', type=str,default='./widerface' ,help='Path for dataset,default WIDERFACE')
parser.add_argument('--batch', type=int, default=32, help='Batch size')
parser.add_argument('--epochs', type=int, default=121, help='Max training epochs')
parser.add_argument('--shuffle', type=bool, default=True, help='Shuffle dataset or not')
parser.add_argument('--img_size', type=int, default=640, help='Input image size')
parser.add_argument('--verbose', type=int, default=20, help='Log verbose')
parser.add_argument('--save_step', type=int, default=10, help='Save every save_step epochs')
parser.add_argument('--eval_step', type=int, default=10, help='Evaluate every eval_step epochs')
parser.add_argument('--save_path', type=str, default='./out', help='Model save path')
parser.add_argument('--training', help='the training mode or not ( True for Training, False for eval', type=bool, default=True)
args = parser.parse_args()
print(args)
return args
def main():
args = get_args()
if not os.path.exists(args.save_path):
os.mkdir(args.save_path)
log_path = os.path.join(args.save_path,'log')
if not os.path.exists(log_path):
os.mkdir(log_path)
data_path = args.data_path
# dataset_train = TrainDataset(train_path,transform=transforms.Compose([RandomCroper(),()]))
dataset_train = TrainDataset('./widerface/train/label.txt',transform=transforms.Compose([RandomErasing(),RandomFlip(),Rotate(),Color(),Resizer(),PadToSquare()]))
# dataset_train = TrainDataset('./widerface/train/label.txt',transform=transforms.Compose([Resizer(),PadToSquare()]))
dataloader_train = DataLoader(dataset_train, num_workers=8, batch_size=args.batch, collate_fn=collater,shuffle=True)
# dataset_val = ValDataset(val_path,transform=transforms.Compose([RandomCroper()]))
dataset_val = TrainDataset('./widerface/train/label.txt',transform=transforms.Compose([Resizer(640),PadToSquare()]))
dataloader_val = DataLoader(dataset_val, num_workers=8, batch_size=args.batch, collate_fn=collater)
total_batch = len(dataloader_train)
# Create torchvision model
return_layers = {'layer2':1,'layer3':2,'layer4':3}
retinaface = torchvision_model.create_retinaface(return_layers)
retinaface_ = retinaface.cuda()
retinaface = torch.nn.DataParallel(retinaface_).cuda()
retinaface.training = True
base_lr=1e-7
# pre_train = torch.load('network.torch')
# cur=retinaface.state_dict()
# for k, v in cur.items():
# if k[12:] in pre_train:
# print(k[12:])
# cur[k]=pre_train[k[12:]]
# retinaface.load_state_dict(cur)
retinaface.load_state_dict(torch.load("/versa/elvishelvis/RetinaYang/out/stage_5_68_full_model_epoch_121.pt"))
lr=base_lr
# optimizer=torch.optim.Adam(retinaface.parameters(),lr=lr)
# fix encoder
for name, value in retinaface.named_parameters():
if 'Landmark' in name:
value.requires_grad = False
lr_cos = lambda n: 0.5 * (1 + np.cos((n) / (args.epochs) * np.pi)) * base_lr
params = filter(lambda p: p.requires_grad==True, retinaface.parameters())
body=filter(lambda p: p.requires_grad==False, retinaface.parameters())
optimizer = torch.optim.Adam([
{'params': body, 'lr': lr*3},
{'params': params, 'lr': lr}]
)
#evaluation the current model
if (args.training==False):
print("not pretrain")
recall, precision, landmakr,miss= eval_widerface.evaluate(dataloader_val,retinaface)
print('Recall:',recall)
print('Precision:',precision)
print("landmark: ",str(landmakr))
print("miss: "+ str(miss))
return
##
print('Start to train.')
epoch_loss = []
iteration = 0
retinaface=retinaface.cuda()
for epoch in range(args.epochs):
lr=lr_cos(epoch)
retinaface.train()
# Training
for iter_num,data in enumerate(dataloader_train):
optimizer.zero_grad()
classification_loss, bbox_regression_loss,ldm_regression_loss = retinaface([data['img'].cuda().float(), data['annot']])
classification_loss = classification_loss.mean()
bbox_regression_loss = bbox_regression_loss.mean()
ldm_regression_loss = ldm_regression_loss.mean()
# loss = classification_loss + 1.0 * bbox_regression_loss + 0.5 * ldm_regression_loss
loss = classification_loss + 0.15*bbox_regression_loss + 0.25*ldm_regression_loss
loss.backward()
optimizer.step()
if iter_num % args.verbose == 0:
log_str = "\n---- [Epoch %d/%d, Batch %d/%d] ----\n" % (epoch, args.epochs, iter_num, total_batch)
table_data = [
['loss name','value'],
['total_loss',str(loss.item())],
['classification',str(classification_loss.item())],
['bbox',str(bbox_regression_loss.item())],
['landmarks',str(ldm_regression_loss.item())]
]
table = AsciiTable(table_data)
log_str +=table.table
print(log_str)
iteration +=1
# Eval
if epoch % args.eval_step == 0:
with open("aaa.txt", 'a') as f:
f.write('-------- RetinaFace Pytorch --------'+'\n')
f.write ('Evaluating epoch {}'.format(epoch)+'\n')
f.write('total_loss:'+str(loss.item())+'\n')
f.write('classification'+str(classification_loss.item())+'\n')
f.write('bbox'+str(bbox_regression_loss.item())+'\n')
f.write('landmarks'+str(ldm_regression_loss.item())+'\n')
f.close()
print('-------- RetinaFace Pytorch --------')
print ('Evaluating epoch {}'.format(epoch))
recall, precision, landmakr,miss= eval_widerface.evaluate(dataloader_val,retinaface)
print('Recall:',recall)
print('Precision:',precision)
print("landmark: ",str(landmakr))
print("miss: "+ str(miss))
with open("aaa.txt", 'a') as f:
f.write('-------- RetinaFace Pytorch --------(not pretrain)'+'\n')
f.write ('Evaluating epoch {}'.format(epoch)+'\n')
f.write('Recall:'+str(recall)+'\n')
f.write('Precision:'+str(precision)+'\n')
f.write("landmark: "+str(landmakr)+'\n')
f.write("miss: "+ str(miss)+'\n')
f.close()
# Save model
if (epoch) % args.save_step == 0:
torch.save(retinaface.state_dict(), args.save_path + '/stage_5_68_full_model_epoch_{}.pt'.format(epoch + 1))
# writer.close()
if __name__=='__main__':
main()