-
Notifications
You must be signed in to change notification settings - Fork 0
/
trainData.py
199 lines (167 loc) · 7.26 KB
/
trainData.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
from data import *
from layers.img_utils import show_loss
from resnet_ssd import build_resnet_ssd
from utils.augmentations import SSDAugmentation
from layers.modules import MultiBoxLoss
from ssd import build_ssd
import os
import time
import torch
from torch.autograd import Variable
import torch.nn as nn
import torch.optim as optim
import torch.backends.cudnn as cudnn
import torch.nn.init as init
import torch.utils.data as data
import warnings
warnings.filterwarnings("ignore")
os.environ['CUDA_ENABLE_DEVICES'] = '0'
class TrainData:
def __init__(self, iter_num, dataset='VOC', dataset_root=VOC_ROOT,
basenet='vgg16_reducedfc.pth', batch_size=4, cuda=False,
lr=1e-4, momentum=0.9, weight_decay=5e-4, gamma=0.1,
save_folder="E:\Code\Examples\Gao\ssd.pytorch-master\weights/"):
self.dataset_name = dataset
self.dataset_root = dataset_root
self.basenet = basenet
self.iter_num = iter_num
self.batch_size = batch_size
self.cuda = cuda
self.lr = lr
self.momentum = momentum
self.weight_decay = weight_decay
self.gamma = gamma
self.save_folder = save_folder
def set_equipment(self):
if torch.cuda.is_available():
print(torch.cuda.get_device_name(0))
if self.cuda:
torch.set_default_tensor_type('torch.cuda.FloatTensor')
if not self.cuda:
print("WARNING: It looks like you have a CUDA device, but aren't " +
"using CUDA.\nRun with --cuda for optimal training speed.")
torch.set_default_tensor_type('torch.FloatTensor')
else:
torch.set_default_tensor_type('torch.FloatTensor')
if not os.path.exists(self.save_folder):
os.mkdir(self.save_folder)
def set_dataset(self):
global cfg, dataset
if self.dataset_name == 'COCO':
if self.dataset_root == VOC_ROOT:
print("WARNING: Using default COCO dataset_root because " +
"--dataset_root was not specified.")
self.dataset_root = COCO_ROOT
cfg = coco
dataset = COCODetection(root=self.dataset_root,
transform=SSDAugmentation(cfg['min_dim'],
MEANS))
elif self.dataset_name == 'VOC':
cfg = voc
dataset = VOCDetection(root=self.dataset_root,
transform=SSDAugmentation(cfg['min_dim'],
MEANS))
return cfg, dataset
def set_net(self):
global ssd_net, net
if self.basenet == 'vgg16_reducedfc.pth':
print('Loading base network vgg16...')
ssd_net = build_ssd('train', cfg['min_dim'], cfg['num_classes'])
net = ssd_net
if self.cuda:
net = torch.nn.DataParallel(ssd_net)
cudnn.benchmark = False
vgg_weights = torch.load(self.save_folder + self.basenet)
ssd_net.vgg.load_state_dict(vgg_weights)
elif self.basenet == 'resnet50-19c8e357.pth':
print('Loading base network resnet50...')
ssd_net = build_resnet_ssd('train', 300, num_classes=21)
net = ssd_net
# 为resnet_ssd加载预训练权重文件
# 1、加载模型
resnet_pretrained_weights = torch.load(self.save_folder + self.basenet)
# 2、初始化网络
resnet_dict = ssd_net.state_dict()
# 3、剔除掉网络中没有的键
pretrained_dict_l = {k: v for k, v in resnet_pretrained_weights.items() if k in resnet_dict}
# 4、用预训练的权重文件，更新网络权重
resnet_dict.update(pretrained_dict_l)
# 5、将更新了的参数放入网络中
ssd_net.load_state_dict(resnet_dict)
return ssd_net, net
def train(self):
self.set_equipment()
cfg, dataset = self.set_dataset()
ssd_net, net = self.set_net()
optimizer = optim.SGD(net.parameters(), lr=self.lr, momentum=self.momentum,
weight_decay=self.weight_decay)
criterion = MultiBoxLoss(cfg['num_classes'], 0.5, True, 0, True, 3, 0.5,
False, self.cuda)
net.train()
# loss counters
loc_loss = 0
conf_loss = 0
epoch = 0
loss_ass = []
print('Loading the dataset...')
epoch_size = len(dataset) // self.batch_size
print('Training SSD on:', dataset.name)
step_index = 0
data_loader = data.DataLoader(dataset, self.batch_size,
shuffle=True, collate_fn=detection_collate,
pin_memory=True)
# create batch iterator
batch_iterator = iter(data_loader)
for iteration in range(self.iter_num):
if iteration in cfg['lr_steps']:
step_index += 1
self.adjust_learning_rate(optimizer, self.gamma, step_index)
# load train data
try:
images, targets = next(batch_iterator)
except StopIteration as e:
batch_iterator = iter(data_loader)
images, targets = next(batch_iterator)
if self.cuda:
images = Variable(images.cuda())
targets = [Variable(ann.cuda(), volatile=True) for ann in targets]
else:
images = Variable(images)
targets = [Variable(ann, volatile=True) for ann in targets]
# forward
t0 = time.time()
out = net(images)
# backprop
optimizer.zero_grad()
loss_l, loss_c = criterion(out, targets)
loss = loss_l + loss_c
loss.backward()
optimizer.step()
t1 = time.time()
loc_loss += loss_l.item()
conf_loss += loss_c.item()
if iteration % 2 == 0:
print('timer: %.4f sec.' % (t1 - t0))
print('iter ' + repr(iteration) + ' || Loss: %.4f ||' % (loss.item()), end=' ')
loss_ass = loss_ass + [loss.item()]
if iteration != 0 and iteration % 5 == 0:
print('Saving state, iter:', iteration)
# torch.save(ssd_net.state_dict(), 'E:\Code\Examples\Gao\ssd.pytorch-master\weights/ssd300_VOC_' +
# repr(iteration) + '.pth')
torch.save(ssd_net.state_dict(), self.save_folder + '' + self.dataset_name + '.pth')
return loss_ass
def adjust_learning_rate(self, optimizer, gamma, step):
"""Sets the learning rate to the initial LR decayed by 10 at every
specified step
# Adapted from PyTorch Imagenet example:
# https://github.com/pytorch/examples/blob/master/imagenet/main.py
"""
lr = self.lr * (gamma ** (step))
for param_group in optimizer.param_groups:
param_group['lr'] = lr
def xavier(param):
init.xavier_uniform(param)
def weights_init(self, m):
if isinstance(m, nn.Conv2d):
self.xavier(m.weight.data)
m.bias.data.zero_()