-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
303 lines (263 loc) · 14.6 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
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
# Copyright 2021 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""train"""
import ast
import argparse
import os
import time
import mindspore.nn as nn
import mindspore
from mindspore import Tensor
from mindspore import context
from mindspore.common import set_seed
from mindspore.context import ParallelMode
from mindspore.communication.management import init, get_rank
from mindspore.train.model import Model
from mindspore.train.serialization import load_checkpoint, load_param_into_net
from mindspore.train.callback import ModelCheckpoint, CheckpointConfig, LossMonitor, TimeMonitor
from src.save_callback import SaveCallback
from src.config import common_config, VehicleNet_train, VeRi_train, VeRi_test
from src.dataset import data_to_mindrecord, create_vehiclenet_dataset
from src.VehicleNet_resnet50 import VehicleNet
from src.lr_generator import lr_steps, lr_steps_2
set_seed(1)
def get_base_param(load_ckpt_path):
"""filter parameters"""
par_dict = load_checkpoint(load_ckpt_path)
new_params_dict = {}
for name in par_dict:
if 'classifier' not in name:
new_params_dict[name] = par_dict[name]
return new_params_dict
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='VehicleNet train.')
parser.add_argument("--run_distribute", type=ast.literal_eval, default=False,
help="Run distribute, default is false.")
parser.add_argument('--device_id', type=int, default=None,
help='device id of GPU or Ascend. (Default: None)')
parser.add_argument('--device_num', type=int, default=1, help='Number of device.')
parser.add_argument('--is_modelarts', type=ast.literal_eval, default=False, help='Train in Modelarts.')
parser.add_argument('--eval_training', type=ast.literal_eval, default=True, help='Eval in training.')
parser.add_argument('--ckpt_url', type=str, default=None, help='Checkpoint file path')
parser.add_argument('--data_url', default=None, help='Location of data.')
parser.add_argument('--train_url', default=None, help='Location of training outputs.')
args_opt = parser.parse_args()
cfg = common_config
VehicleNet_cfg = VehicleNet_train
VeRi_cfg = VeRi_train
VeRi_test_cfg = VeRi_test
device_target = cfg.device_target
context.set_context(mode=context.GRAPH_MODE, device_target=device_target)
if args_opt.run_distribute:
if device_target == "Ascend":
device_id = int(os.getenv('DEVICE_ID'))
context.set_context(device_id=device_id, enable_auto_mixed_precision=True)
context.set_auto_parallel_context(parallel_mode=ParallelMode.DATA_PARALLEL,
gradients_mean=True,
auto_parallel_search_mode="recursive_programming")
init()
else:
raise ValueError("Unsupported platform.")
else:
if device_target == "Ascend":
if args_opt.device_id is not None:
context.set_context(device_id=args_opt.device_id)
else:
context.set_context(device_id=cfg.device_id)
else:
raise ValueError("Unsupported platform.")
train_dataset_path = cfg.dataset_path
pre_trained_file = cfg.pre_trained_file
if args_opt.is_modelarts:
import moxing as mox
mox.file.copy_parallel(src_url=args_opt.data_url,
dst_url='/cache/dataset_train/device_' + os.getenv('DEVICE_ID'))
zip_command = "unzip -o /cache/dataset_train/device_" + os.getenv('DEVICE_ID') \
+ "/VehicleNet_mindrecord_v3.zip -d /cache/dataset_train/device_" + os.getenv('DEVICE_ID')
os.system(zip_command)
train_dataset_path = '/cache/dataset_train/device_' + os.getenv('DEVICE_ID') + '/VehicleNet/'
pre_trained_file = '/cache/dataset_train/device_' + os.getenv('DEVICE_ID') + '/resnet50.ckpt'
mindrecord_dir = cfg.mindrecord_dir
prefix = "first_train_VehicleNet.mindrecord"
mindrecord_file = os.path.join(mindrecord_dir, prefix)
if not os.path.exists(mindrecord_file):
if not os.path.isdir(mindrecord_dir):
os.makedirs(mindrecord_dir)
print("Create mindrecord for train.")
data_to_mindrecord(train_dataset_path, True, True, False, mindrecord_file)
print("Create mindrecord done, at {}".format(mindrecord_dir))
while not os.path.exists(mindrecord_file + ".db"):
time.sleep(5)
dataset = create_vehiclenet_dataset(mindrecord_file, batch_size=VehicleNet_cfg.batch_size,
device_num=args_opt.device_num, is_training=True)
step_per_epoch_first = dataset.get_dataset_size()
mindrecord_dir = cfg.mindrecord_dir
prefix = "test_VehicleNet.mindrecord"
test_mindrecord_file = os.path.join(mindrecord_dir, prefix)
if not os.path.exists(test_mindrecord_file):
if not os.path.isdir(mindrecord_dir):
os.makedirs(mindrecord_dir)
print("Create mindrecord for test.")
data_to_mindrecord(eval_dataset_path, False, False, True, test_mindrecord_file)
print("Create mindrecord done, at {}".format(mindrecord_dir))
while not os.path.exists(test_mindrecord_file + ".db"):
time.sleep(5)
prefix = "query_VehicleNet.mindrecord"
query_mindrecord_file = os.path.join(mindrecord_dir, prefix)
if not os.path.exists(query_mindrecord_file):
if not os.path.isdir(mindrecord_dir):
os.makedirs(mindrecord_dir)
print("Create mindrecord for query.")
data_to_mindrecord(eval_dataset_path, False, False, False, query_mindrecord_file)
print("Create mindrecord done, at {}".format(mindrecord_dir))
while not os.path.exists(query_mindrecord_file + ".db"):
time.sleep(5)
test_dataset = create_vehiclenet_dataset(test_mindrecord_file, batch_size=1, device_num=1, is_training=False)
query_dataset = create_vehiclenet_dataset(query_mindrecord_file, batch_size=1, device_num=1, is_training=False)
test_data_num = test_dataset.get_dataset_size()
query_data_num = query_dataset.get_dataset_size()
net = VehicleNet(class_num=VehicleNet_cfg.num_classes)
net_test = VehicleNet(class_num=VeRi_test_cfg.num_classes)
net_test.classifier.classifier = nn.SequentialCell()
if cfg.pre_trained:
param_dict = load_checkpoint(pre_trained_file)
load_param_into_net(net, param_dict)
loss = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction='mean')
lr_base = lr_steps(0.1 * VehicleNet_cfg.lr_init, VehicleNet_cfg.epoch_size, step_per_epoch_first)
lr_ignored = lr_steps(VehicleNet_cfg.lr_init, VehicleNet_cfg.epoch_size, step_per_epoch_first)
base_params = list(filter(lambda x: 'classifier' not in x.name, net.trainable_params()))
ignored_params = list(filter(lambda x: 'classifier' in x.name, net.trainable_params()))
train_params = [{'params': base_params, 'lr': Tensor(lr_base, mindspore.float32)},
{'params': ignored_params, 'lr': Tensor(lr_ignored, mindspore.float32)}]
opt = nn.SGD(train_params, momentum=VehicleNet_cfg.momentum, weight_decay=VehicleNet_cfg.weight_decay)
model = Model(net, loss_fn=loss, optimizer=opt)
time_cb = TimeMonitor(data_size=step_per_epoch_first)
loss_cb = LossMonitor()
cb = [time_cb, loss_cb]
if cfg.save_checkpoint:
config_ck = CheckpointConfig(save_checkpoint_steps=cfg.save_checkpoint_epochs * step_per_epoch_first,
keep_checkpoint_max=cfg.keep_checkpoint_max)
if args_opt.is_modelarts:
save_checkpoint_path = '/cache/train_output/checkpoint'
if args_opt.device_num == 1:
ckpt_cb = ModelCheckpoint(prefix='first_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
if args_opt.device_num > 1 and get_rank() % 8 == 0:
if args_opt.eval_training:
save_cb = SaveCallback(net_test, test_dataset, query_dataset, 10, VeRi_test_cfg)
cb.append(save_cb)
ckpt_cb = ModelCheckpoint(prefix='first_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
else:
save_checkpoint_path = cfg.checkpoint_dir
if not os.path.isdir(save_checkpoint_path):
os.makedirs(save_checkpoint_path)
if args_opt.device_num == 1:
if args_opt.eval_training:
save_cb = SaveCallback(net_test, test_dataset, query_dataset, 10, VeRi_test_cfg)
cb.append(save_cb)
ckpt_cb = ModelCheckpoint(prefix='first_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
if args_opt.device_num > 1 and get_rank() % 8 == 0:
if args_opt.eval_training:
save_cb = SaveCallback(net_test, test_dataset, query_dataset, 10, VeRi_test_cfg)
cb.append(save_cb)
ckpt_cb = ModelCheckpoint(prefix='first_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
model.train(VehicleNet_cfg.epoch_size, dataset, callbacks=cb)
time.sleep(120)
if args_opt.is_modelarts:
mox.file.copy_parallel(src_url='/cache/train_output', dst_url=args_opt.train_url)
mindrecord_dir = cfg.mindrecord_dir
prefix = "second_train_VehicleNet.mindrecord"
mindrecord_file = os.path.join(mindrecord_dir, prefix)
if not os.path.exists(mindrecord_file):
if not os.path.isdir(mindrecord_dir):
os.makedirs(mindrecord_dir)
print("Create mindrecord for train.")
data_to_mindrecord(train_dataset_path, True, False, False, mindrecord_file)
print("Create mindrecord done, at {}".format(mindrecord_dir))
while not os.path.exists(mindrecord_file + ".db"):
time.sleep(5)
dataset = create_vehiclenet_dataset(mindrecord_file, batch_size=VeRi_cfg.batch_size,
device_num=args_opt.device_num, is_training=True)
step_per_epoch_second = dataset.get_dataset_size()
net = VehicleNet(class_num=VeRi_cfg.num_classes)
first_trained_file = '/cache/train_output/checkpoint/first_train_vehiclenet-80_' + \
str(step_per_epoch_first) + '.ckpt'
# first_trained_file = '../../checkpoint/first_train_vehiclenet-80_' + str(step_per_epoch_first) + '.ckpt'
param_dict = get_base_param(first_trained_file)
load_param_into_net(net, param_dict)
loss = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction='mean')
lr_base = lr_steps_2(0.1 * VeRi_cfg.lr_init, VeRi_cfg.epoch_size, step_per_epoch_second)
lr_ignored = lr_steps_2(VeRi_cfg.lr_init, VeRi_cfg.epoch_size, step_per_epoch_second)
base_params = list(filter(lambda x: 'classifier' not in x.name, net.trainable_params()))
ignored_params = list(filter(lambda x: 'classifier' in x.name, net.trainable_params()))
train_params = [{'params': base_params, 'lr': Tensor(lr_base, mindspore.float32)},
{'params': ignored_params, 'lr': Tensor(lr_ignored, mindspore.float32)}]
opt = nn.SGD(train_params, momentum=VeRi_cfg.momentum, weight_decay=VeRi_cfg.weight_decay)
model = Model(net, loss_fn=loss, optimizer=opt)
time_cb = TimeMonitor(data_size=step_per_epoch_second)
loss_cb = LossMonitor()
cb = [time_cb, loss_cb]
if cfg.save_checkpoint:
config_ck = CheckpointConfig(save_checkpoint_steps=cfg.save_checkpoint_epochs * step_per_epoch_second,
keep_checkpoint_max=cfg.keep_checkpoint_max)
if args_opt.is_modelarts:
save_checkpoint_path = '/cache/train_output/checkpoint'
if args_opt.device_num == 1:
ckpt_cb = ModelCheckpoint(prefix='second_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
if args_opt.device_num > 1 and get_rank() % 8 == 0:
if args_opt.eval_training:
save_cb = SaveCallback(net_test, test_dataset, query_dataset, 10, VeRi_test_cfg)
cb.append(save_cb)
ckpt_cb = ModelCheckpoint(prefix='second_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
else:
save_checkpoint_path = cfg.checkpoint_dir
if not os.path.isdir(save_checkpoint_path):
os.makedirs(save_checkpoint_path)
if args_opt.device_num == 1:
if args_opt.eval_training:
save_cb = SaveCallback(net_test, test_dataset, query_dataset, 5, VeRi_test_cfg)
cb.append(save_cb)
ckpt_cb = ModelCheckpoint(prefix='second_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
if args_opt.device_num > 1 and get_rank() % 8 == 0:
if args_opt.eval_training:
save_cb = SaveCallback(net_test, test_dataset, query_dataset, 10, VeRi_test_cfg)
cb.append(save_cb)
ckpt_cb = ModelCheckpoint(prefix='second_train_vehiclenet',
directory=save_checkpoint_path,
config=config_ck)
cb += [ckpt_cb]
model.train(VeRi_cfg.epoch_size, dataset, callbacks=cb)
if args_opt.is_modelarts:
mox.file.copy_parallel(src_url='/cache/train_output', dst_url=args_opt.train_url)