forked from mitmul/chainer-cifar10
-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
212 lines (177 loc) · 7.28 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
#!/usr/bin/env python
# -*- coding: utf-8 -*-
from chainer import computational_graph
from chainer import cuda
from chainer import optimizers
from chainer import serializers
from chainer import Variable
from dataset import load_dataset
from multiprocessing import Process
from multiprocessing import Queue
from transform import Transform
import argparse
import chainer
import draw_loss
import imp
import logging
import numpy as np
import os
import shutil
import six
import time
def create_result_dir(args):
result_dir = 'results/' + os.path.basename(args.model).split('.')[0]
result_dir += '_' + time.strftime('%Y-%m-%d_%H-%M-%S_')
result_dir += str(time.time()).replace('.', '')
if not os.path.exists(result_dir):
os.makedirs(result_dir)
log_fn = '%s/log.txt' % result_dir
logging.basicConfig(
format='%(asctime)s [%(levelname)s] %(message)s',
filename=log_fn, level=logging.DEBUG)
logging.info(args)
args.log_fn = log_fn
args.result_dir = result_dir
def get_model_optimizer(args):
model_fn = os.path.basename(args.model)
model = imp.load_source(model_fn.split('.')[0], args.model).model
dst = '%s/%s' % (args.result_dir, model_fn)
if not os.path.exists(dst):
shutil.copy(args.model, dst)
dst = '%s/%s' % (args.result_dir, os.path.basename(__file__))
if not os.path.exists(dst):
shutil.copy(__file__, dst)
# prepare model
if args.gpu >= 0:
cuda.get_device(args.gpu).use()
model.to_gpu()
# prepare optimizer
if 'opt' in args:
# prepare optimizer
if args.opt == 'MomentumSGD':
optimizer = optimizers.MomentumSGD(lr=args.lr, momentum=0.9)
elif args.opt == 'Adam':
optimizer = optimizers.Adam(alpha=args.alpha)
elif args.opt == 'AdaGrad':
optimizer = optimizers.AdaGrad(lr=args.lr)
else:
raise Exception('No optimizer is selected')
optimizer.setup(model)
if args.opt == 'MomentumSGD':
optimizer.add_hook(
chainer.optimizer.WeightDecay(args.weight_decay))
return model, optimizer
else:
print('No optimizer generated.')
return model
def augmentation(args, aug_queue, data, label, train):
trans = Transform(args)
np.random.seed(int(time.time()))
perm = np.random.permutation(data.shape[0])
if train:
for i in six.moves.range(0, data.shape[0], args.batchsize):
chosen_ids = perm[i:i + args.batchsize]
x = np.asarray(data[chosen_ids], dtype=np.float32)
x = x.transpose((0, 3, 1, 2))
t = np.asarray(label[chosen_ids], dtype=np.int32)
aug_queue.put((x, t))
else:
for i in six.moves.range(data.shape[0]):
aug = trans(data[i])
x = np.asarray(aug, dtype=np.float32).transpose((0, 3, 1, 2))
t = np.asarray(np.repeat(label[i], len(aug)), dtype=np.int32)
aug_queue.put((x, t))
aug_queue.put(None)
return
def one_epoch(args, model, optimizer, data, label, epoch, train):
model.train = train
xp = cuda.cupy if args.gpu >= 0 else np
# for parallel augmentation
aug_queue = Queue()
aug_worker = Process(target=augmentation,
args=(args, aug_queue, data, label, train))
aug_worker.start()
logging.info('data loading started')
sum_accuracy = 0
sum_loss = 0
num = 0
while True:
datum = aug_queue.get()
if datum is None:
break
x, t = datum
volatile = 'off' if train else 'on'
x = Variable(xp.asarray(x), volatile=volatile)
t = Variable(xp.asarray(t), volatile=volatile)
if train:
optimizer.update(model, x, t)
if epoch == 1 and num == 0:
with open('{}/graph.dot'.format(args.result_dir), 'w') as o:
g = computational_graph.build_computational_graph(
(model.loss, ), remove_split=True)
o.write(g.dump())
sum_loss += float(model.loss.data) * t.data.shape[0]
sum_accuracy += float(model.accuracy.data) * t.data.shape[0]
num += t.data.shape[0]
logging.info('{:05d}/{:05d}\tloss:{:.3f}\tacc:{:.3f}'.format(
num, data.shape[0], sum_loss / num, sum_accuracy / num))
else:
pred = model(x, t).data
pred = pred.mean(axis=0)
acc = int(pred.argmax() == t.data[0])
sum_accuracy += acc
num += 1
logging.info('{:05d}/{:05d}\tacc:{:.3f}'.format(
num, data.shape[0], sum_accuracy / num))
del x, t
if train and (epoch == 1 or epoch % args.snapshot == 0):
model_fn = '{}/epoch-{}.model'.format(args.result_dir, epoch)
opt_fn = '{}/epoch-{}.state'.format(args.result_dir, epoch)
serializers.save_hdf5(model_fn, model)
serializers.save_hdf5(opt_fn, optimizer)
if train:
logging.info('epoch:{}\ttrain loss:{}\ttrain accuracy:{}'.format(
epoch, sum_loss / data.shape[0], sum_accuracy / data.shape[0]))
else:
logging.info('epoch:{}\ttest loss:{}\ttest accuracy:{}'.format(
epoch, sum_loss / data.shape[0], sum_accuracy / data.shape[0]))
draw_loss.draw_loss_curve('{}/log.txt'.format(args.result_dir),
'{}/log.png'.format(args.result_dir), epoch)
aug_worker.join()
logging.info('data loading finished')
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--model', type=str, default='models/VGG.py')
parser.add_argument('--gpu', type=int, default=0)
parser.add_argument('--epoch', type=int, default=100)
parser.add_argument('--batchsize', type=int, default=128)
parser.add_argument('--snapshot', type=int, default=10)
parser.add_argument('--datadir', type=str, default='data')
# optimization
parser.add_argument('--opt', type=str, default='MomentumSGD',
choices=['MomentumSGD', 'Adam', 'AdaGrad'])
parser.add_argument('--weight_decay', type=float, default=0.0001)
parser.add_argument('--alpha', type=float, default=0.001)
parser.add_argument('--lr', type=float, default=0.01)
parser.add_argument('--lr_decay_freq', type=int, default=5)
parser.add_argument('--lr_decay_ratio', type=float, default=0.1)
parser.add_argument('--validate_freq', type=int, default=1)
parser.add_argument('--seed', type=int, default=1701)
args = parser.parse_args()
np.random.seed(args.seed)
# os.environ['CHAINER_TYPE_CHECK'] = str(args.type_check)
# os.environ['CHAINER_SEED'] = str(args.seed)
# create result dir
create_result_dir(args)
# create model and optimizer
model, optimizer = get_model_optimizer(args)
dataset = load_dataset(args.datadir)
tr_data, tr_labels, te_data, te_labels = dataset
# learning loop
for epoch in range(1, args.epoch + 1):
logging.info('learning rate:{}'.format(optimizer.lr))
one_epoch(args, model, optimizer, tr_data, tr_labels, epoch, True)
if epoch == 1 or epoch % args.validate_freq == 0:
one_epoch(args, model, optimizer, te_data, te_labels, epoch, False)
if args.opt == 'MomentumSGD' and epoch % args.lr_decay_freq == 0:
optimizer.lr *= args.lr_decay_ratio