-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtrain.py
173 lines (140 loc) · 6.04 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
import os
import torch
import torch.nn as nn
import torch.optim as optim
import torch.backends.cudnn as cudnnz
import torch.nn.init as init
import argparse
from torch.autograd import Variable
import torch.utils.data as data
from data import *
from utils.augmentations import SSDAugmentation,StixelAugmentation
from layers.modules import MultiBoxLoss,StixelLoss
from StixelNet import build_net
import numpy as np
import time
parser = argparse.ArgumentParser()
parser.add_argument('--batch_size', default=4, type=int, help='Batch size for training')
parser.add_argument('--num_workers', default=0, type=int, help='Number of workers used in dataloading')
parser.add_argument('--lr', '--learning-rate', default=3e-5, type=float, help='initial learning rate')
parser.add_argument('--momentum', default=0.9, type=float, help='momentum')
parser.add_argument('--weight_decay', default=5e-4, type=float, help='Weight decay for SGD')
parser.add_argument('--gamma', default=0.9, type=float, help='Gamma update for SGD')
parser.add_argument('--basepath_d',type=str,help='The basepath of KittiTracking')
parser.add_argument('--basepath_s',type=str,help='The basepath of Kitti Raw Data')
parser.add_argument('--gt_path_s',type=str,help='The path of Stixel Ground Trurh')
parser.add_argument('--resume',type=str,help='The path of checkpoint')
args = parser.parse_args()
torch.set_default_tensor_type('torch.cuda.FloatTensor')
ssd_dim = (800,370) # the size of image after resize (width,height)
means = (104, 117, 123)
num_classes = 9 + 1
batch_size = args.batch_size
def xavier(param):
init.xavier_uniform(param)
def weights_init(m):
if isinstance(m, nn.Conv2d):
xavier(m.weight.data)
m.bias.data.zero_()
if args.resume is None:
net = build_net('train', ssd_dim, num_classes)
vgg_weights = torch.load('weights/vgg16_reducedfc.pth')
print('Loading base network...')
net.vgg.load_state_dict(vgg_weights)
net.extras.apply(weights_init)
net.loc.apply(weights_init)
net.conf.apply(weights_init)
else:
net=torch.load(args.resume)
savename='weights/kitti_%f_%.3f'%(args.lr,args.gamma)
cudnnz.benchmark = True
net = net.cuda()
optimizer = optim.SGD(net.parameters(), lr=args.lr,
momentum=args.momentum, weight_decay=args.weight_decay)
def dection_train():
logfile=open("./log/log_%f_%.3f.txt"%(args.lr,args.gamma),"w")
net.train()
step=0
index=1
criterion = MultiBoxLoss(num_classes, 0.5, True, 0, True, 3, 0.5, False, True)
dataset=KittiTracking(args.basepath,index,SSDAugmentation(size=ssd_dim, mean=means))
data_loader=data.DataLoader(dataset, batch_size, num_workers=args.num_workers,
shuffle=True, collate_fn=detection_collate, pin_memory=True)
for epoch in range(500):
if epoch%10==0:
lr=adjust_learning_rate(optimizer, args.gamma, step)
step = step + 1
print("present learning rate is %.6f"%lr)
loc_loss = 0 # epoch
conf_loss = 0
for i,(images,targets) in enumerate(data_loader):
images = Variable(images.cuda())
targets = [Variable(anno.cuda(), volatile=True) for anno in targets]
# forward
dec, stixel = net(images)
# backprop
optimizer.zero_grad()
loss_l, loss_c = criterion(dec, targets)
loss = loss_l + loss_c
loss.backward()
optimizer.step()
loc_loss += loss_l.data[0]
conf_loss += loss_c.data[0]
if i % 10 == 0:
print("Epoch:%d batch:%d/%d loss:%.4f L(loc):%.4f L(conf):%.4f"%(epoch,i,len(data_loader),loss.data[0],loss_l.data[0],loss_c.data[0]))
loc_loss=loc_loss*batch_size/len(dataset)
conf_loss=conf_loss*batch_size/len(dataset)
print("Totol loss:%.4f L(loc):%.4f L(conf):%.4f"%(loc_loss+conf_loss,loc_loss,conf_loss))
logfile.write("Epoch:%d Totol loss:%.4f L(loc):%.4f L(conf):%.4f\n"%(epoch,loc_loss+conf_loss,loc_loss,conf_loss))
if epoch % 10 == 0 and epoch>0:
print('Saving state, epoch:', epoch)
torch.save(net, savename+('_%d.pth'%epoch))
torch.save(net, savename+('_%d.pth'%epoch))
logfile.close()
def stixel_train():
logfile=open("./log/log_%f_%.3f.txt"%(args.lr,args.gamma),"w")
net.train()
printfrq=10
step=0
dataset = StixelKitti(args.basepath_s,args.gt_path_s,StixelAugmentation(size=ssd_dim, mean=means))
data_loader=data.DataLoader(dataset, batch_size, num_workers=args.num_workers,
shuffle=False, pin_memory=True)
lossfunction=StixelLoss()
minloss=9999
for epoch in range(200):
if epoch%10==0:
lr=adjust_learning_rate(optimizer, args.gamma, step)
step = step + 1
avgloss = 0
for i,(images,havetargets,targets) in enumerate(data_loader):
images=Variable(images).cuda()
havetargets=Variable(havetargets).cuda()
targets=Variable(targets).cuda()
dec , stixel =net(images)
optimizer.zero_grad()
loss=lossfunction(stixel,havetargets,targets)
loss.backward()
optimizer.step()
avgloss=avgloss+loss.data[0]
if i % printfrq == 0:
if i!=0:
avgloss=avgloss/printfrq
print("Epoch: %d batch: %d lr: %.6f loss: %.6f" % (
epoch, i, lr,avgloss))
logfile.write("Epoch: %d batch: %d lr: %.6f loss: %.6f\n" % (
epoch, i, lr,avgloss))
if avgloss < minloss:
minloss = avgloss
avgloss=0
if epoch % 10 == 0 and epoch>0:
torch.save(net, savename+('_%d.pth'%epoch))
torch.save(net, savename+('_%d.pth'%epoch))
logfile.close()
def adjust_learning_rate(optimizer, gamma, step):
lr = args.lr * (gamma ** (step))
for param_group in optimizer.param_groups:
param_group['lr'] = lr
return lr
if __name__ == '__main__':
#dection_train()
stixel_train()