-
Notifications
You must be signed in to change notification settings - Fork 2
/
train.py
251 lines (230 loc) · 10.5 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
import os
import json
import time
import argparse
import torch
import shutil
import torch.nn as nn
from tensorboardX import SummaryWriter
from torch.autograd import Variable
from torchvision import transforms
from src.dataset import *
from src.model import CSRNet
from src.utils import *
parser = argparse.ArgumentParser(description = "PyTorch CSRNet vehicle")
parser.add_argument("--train_json",
default = "/data/wangyf/datasets/TRANCOS_v3/image_sets/training.txt",
help = "path to train json")
parser.add_argument("--val_json",
default = "/data/wangyf/datasets/TRANCOS_v3/image_sets/validation.txt",
help = "path to val json")
parser.add_argument("--gpu", default = '6', type = str, help = "GPU id to use.")
parser.add_argument("--task", default = 's_1_100_0320', type = str,
help = "task id to use.")
parser.add_argument("--pre", '-p', default = None, type = str,
help = "path to pretrained model")
parser.add_argument("--lr", default = 1e-6, type = float, help = "original learning rate") ## TODO
parser.add_argument("--epochs_drop", default = 30, type = int, help = "epochs_drop") ## TODO
parser.add_argument("--start_epoch", default = 0, type = int, help = "start epoch")
parser.add_argument("--epochs", default = 60,type = int, help = "epoch") ## TODO
parser.add_argument("--batch_size", default = 1, type = int, help = "batch size") ## TODO
parser.add_argument("--momentum", default = 0.95, type = float) ## TODO
parser.add_argument("--decay", default = 5 * 1e-4, type = float) ## TODO
parser.add_argument("--workers", default = 4,type = int) ## TODO
parser.add_argument("--print_freq", default = 40, type = int, help = "show train log information")
args = parser.parse_args()
os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
try:
from termcolor import cprint
except ImportError:
cprint = None
## create log for ssh check
localtime = time.strftime("%Y-%m-%d", time.localtime())
log_file = open("./logs/record" + localtime + ".txt", 'w')
## create tensorboard dir
logDir = "./tblogs/0320"
if os.path.exists(logDir):
shutil.rmtree(logDir) # remove recursive dir
writer = SummaryWriter(logDir)
def log_print(text, color = None, on_color = None,
attrs = None, log_file = log_file):
print(text, file = log_file)
if cprint is not None:
cprint(text, color = color, on_color = on_color, attrs = attrs)
else:
print(text)
# .txt(str) convert to list, list is the image path
train_list = [] # 403
val_list = [] # 420
with open(args.train_json, 'r') as f1:
for line in f1.readlines():
line = line.strip("\n")
train_list.append(line)
with open(args.val_json, 'r') as f2:
for line in f2.readlines():
line = line.strip("\n")
val_list.append(line)
# Sets the learning rate to the initial LR decayed by 10 every 30 epochs
def adjust_learning_rate(optimizer, epoch):
factor = 0.1 ** (epoch // args.epochs_drop)
args.lr = args.lr * factor
# print(len(optimizer.param_groups)) 1
for param_group in optimizer.param_groups:
param_group['lr'] = args.lr
### Computes and stores the average and current value
class AverageMeter(object):
def __init__(self):
self.reset()
def reset(self):
self.cur = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, cur, n = 1):
self.cur = cur
self.sum += cur * n
self.count += n
self.avg = self.sum / self.count
### training
def train(train_list, model, criterion, optimizer, epoch):
'''
losses: batch_size loss value, include cur and avg
batch_time: batch_size train time, include cur and avg
data_time: batch_size loader input time, include cur and avg
'''
losses = AverageMeter()
batch_time = AverageMeter()
data_time = AverageMeter()
mean = [0.485, 0.456, 0.406]
std = [0.229, 0.224, 0.225]
transform_train = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize(mean, std)
])
DatasetLoader_train = listDataset(train_list,
shuffle = True,
transform = transform_train,
train = True,
seen = model.seen,
batch_size = args.batch_size,
num_workers = args.workers)
train_loader = torch.utils.data.DataLoader(DatasetLoader_train,
batch_size = args.batch_size)
log_text = "epoch %d, processed %d samples, lr % .10f "\
%(epoch, epoch * len(train_loader.dataset), args.lr)
log_print(log_text, color = "green", attrs = ["bold"])
model.train()
end = time.time()
for i,(img, target)in enumerate(train_loader):
'''img: ; target: torch.Size([1, H/8, W/8])'''
data_time.update(time.time()- end)
img = img.cuda() # torch.Size([batch_size,3,H,W])
img = Variable(img) # torch.FloatTensor to Variable
output = model(img) # torch.Size([batch_size,1,H/8,W/8])
target = target.type(torch.FloatTensor).unsqueeze(1).cuda() # torch.Size([1,batch_size,H/8,W/8])
target = Variable(target)
loss = criterion(output, target)
losses.update(loss.item(), img.size(0)) # img.size[0] == batch_size
optimizer.zero_grad()
loss.backward()
optimizer.step()
batch_time.update(time.time() - end)
end = time.time()
if i % args.print_freq == 0:
print_str = "Epoch: [{0}][{1}/{2}]\t" \
.format(epoch, i, len(train_loader))
print_str += "Data time {data_time.cur:.3f}({data_time.avg:.3f})\t" \
.format(data_time=data_time)
print_str += "Batch time {batch_time.cur:.3f}({batch_time.avg:.3f})\t" \
.format(batch_time=batch_time)
print_str += "Loss {loss.cur:.4f}({loss.avg:.4f})\t" \
.format(loss=losses)
# print(print_str)
log_print(print_str, color="green", attrs=["bold"])
return losses.avg
### val
def validate(val_list, model, epoch, criterion):
data_time = AverageMeter()
batch_time = AverageMeter()
losses = AverageMeter()
mean = [0.485, 0.456, 0.406]
std = [0.229, 0.224, 0.225]
transform_val = transforms.Compose([transforms.ToTensor(),
transforms.Normalize(mean, std)])
DatasetLoader_val = listDataset(val_list,
shuffle = False,
transform = transform_val,
train = False)
val_loader = torch.utils.data.DataLoader(DatasetLoader_val,
batch_size = args.batch_size) # [C,H,W]
model.eval()
end = time.time()
mae = 0.0
for i,(img, gt_density_map) in enumerate(val_loader):
data_time.update(time.time() - end)
img = img.cuda()
img = Variable(img)
gt_density_map = gt_density_map.type(torch.FloatTensor).unsqueeze(1)
gt_density_map = gt_density_map.cuda()
gt_density_map = Variable(gt_density_map)
et_density_map = model(img)
loss = criterion(gt_density_map, et_density_map)
losses.update(loss.item(), img.size(0))
batch_time.update(time.time() - end)
mae += abs(et_density_map.data.sum() - gt_density_map.sum()) # todo
end = time.time()
if i % args.print_freq == 0:
print_str = "Epoch: [{0}][{1}/{2}]\t" \
.format(epoch, i, len(val_loader))
print_str += "Data time {data_time.cur:.3f}({data_time.avg:.3f})\t" \
.format(data_time=data_time)
print_str += "Batch time {batch_time.cur:.3f}({batch_time.avg:.3f})\t" \
.format(batch_time=batch_time)
print_str += "Loss {loss.cur:.4f}({loss.avg:.4f})\t" \
.format(loss=losses)
# print(print_str)
log_print(print_str, color="red", attrs=["bold"])
mae = mae / len(val_loader)
return losses.avg, mae
def main():
global best_mae
best_mae = 1e6
seed = time.time()
torch.cuda.manual_seed(seed)
model = CSRNet() # CSRNet[weight, bias] Initialize
model = model.cuda()
criterion = SANetLoss(1).cuda()
optimizer = torch.optim.SGD(model.parameters(),
args.lr,
momentum = args.momentum,
weight_decay = args.decay) ## TODO
## pre-train model
if args.pre:
if os.path.isfile(args.pre):
print("=> loading checkpoint '{}'".format(args.pre))
checkpoint = torch.load(args.pre)
args.start_epoch = checkpoint['epoch']
best_mae = checkpoint['best_mae']
model.load_state_dict(checkpoint['state_dict'])
optimizer.load_state_dict(checkpoint['optimizer'])
print("=> loaded checkpoint '{}' (epoch {})" .format(args.pre, checkpoint['epoch']))
else:
print("=> no checkpoint found at '{}'".format(args.pre))
for epoch in range(args.start_epoch, args.epochs):
train_loss = train(train_list, model, criterion, optimizer, epoch)
val_loss, val_mae = validate(val_list, model, epoch, criterion)
writer.add_scalar("/train_loss", train_loss, epoch)
writer.add_scalar("/val_loss", val_loss, epoch)
is_best = val_mae < best_mae
best_mae = min(val_mae, best_mae)
print(epoch, val_mae, best_mae)
## save model 1-400
save_checkpoint({"epoch": epoch + 1,
"arch": args.pre,
"state_dict": model.state_dict(),
"best_mae": best_mae,
"optimizer": optimizer.state_dict(), },
is_best,
args.task)
if __name__ == '__main__':
main()