-
Notifications
You must be signed in to change notification settings - Fork 5
/
train.py
201 lines (169 loc) · 8.02 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
# -*- coding: utf-8 -*-
# @Time : 2018/8/16 21:29
# @Author : zsz
# @Site :
# @File : train.py
# @Software: PyCharm
# @Desc :
# @license : Copyright(C), Your Company
from __future__ import absolute_import
from __future__ import print_function
from __future__ import division
import time
import libs.nets.network_factory
import libs.data.data_batch as data_batch
import libs.nets.build_fpn as fpn
from configs.train_config import TRIAN_CONFIG
import tensorflow.contrib.slim as slim
import tensorflow as tf
import os
import re
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '1'
tf.app.flags.DEFINE_float(
'gpu_memory_fraction', 0.85, 'GPU memory fraction to use.')
tf.app.flags.DEFINE_string(
'train_dir', './dssd_tfmodel/synth_model/dssd_resnet',
'Directory where checkpoints and event logs are written to.')
tf.app.flags.DEFINE_string(
'dataset_dir', '/home/zsz/datasets/synth-tf/', 'The directory where the dataset files are stored.')
# 'dataset_dir', './tfrecord_for_train/huawei_ano_with_difficult_v2', 'The directory where the dataset files are stored.')
tf.app.flags.DEFINE_integer(
'batch_size', 16, 'The number of samples in each batch.')
tf.app.flags.DEFINE_integer(
'gpu_num', 2, 'the number of gpu use'
)
# =========================================================================== #
tf.app.flags.DEFINE_string(
#'checkpoint_path', 'huawei_model'
'checkpoint_path', 'resnet_model/resnet_v1_101.ckpt',
# 'checkpoint_path', './ssd_model',
# 'checkpoint_path', './dssd_tfmodel/synth_model/huawei_synth_v3_dssd_final_model/',#retrain a new model
'The path to a checkpoint from which to fine-tune.')
tf.app.flags.DEFINE_bool(
'ignore_missing_vars', False,
'The parameter which means ignore missing vars in the checkpoint'
)
FLAGS = tf.app.flags.FLAGS
def tower_loss(scope, images, segmaps):
"""Calculate the total loss on a single tower running the CIFAR model.
Args:
scope: unique prefix string identifying the CIFAR tower, e.g. 'tower_0'
images: Images. 4D tensor of shape [batch_size, height, width, 3].
labels: Labels. 1D tensor of shape [batch_size].
Returns:
Tensor of shape [] containing the total loss for a batch of data
"""
# Build inference Graph.
# FPN = fpn.FPN()
# fpn_net = FPN.build_fpn()
# print(fpn_net)
fpn_model = fpn.FPN(TRIAN_CONFIG['net_name'], images, is_training=True)
loss = fpn_model.build_loss(segmaps)
losses = tf.get_collection('losses', scope)
total_loss = tf.add_n(losses, name='total_loss')
for l in losses + [total_loss]:
loss_name = re.sub('%s_[0-9]*/' % 'psenet_tower', '', l.op.name)
tf.summary.scalar(loss_name, l)
return total_loss
def average_gradients(tower_grads):
"""Calculate the average gradient for each shared variable across all towers.
Note that this function provides a synchronization point across all towers.
Args:
tower_grads: List of lists of (gradient, variable) tuples. The outer list
is over individual gradients. The inner list is over the gradient
calculation for each tower.
Returns:
List of pairs of (gradient, variable) where the gradient has been averaged
across all towers.
"""
average_grads = []
for grad_and_vars in zip(*tower_grads):
# Note that each grad_and_vars looks like the following:
# ((grad0_gpu0, var0_gpu0), ... , (grad0_gpuN, var0_gpuN))
grads = []
for g, _ in grad_and_vars:
# Add 0 dimension to the gradients to represent the tower.
expanded_g = tf.expand_dims(g, 0)
# Append on a 'tower' dimension which we will average over below.
grads.append(expanded_g)
# Average over the 'tower' dimension.
grad = tf.concat(axis=0, values=grads)
grad = tf.reduce_mean(grad, 0)
# Keep in mind that the Variables are redundant because they are shared
# across towers. So .. we will just return the first tower's pointer to
# the Variable.
v = grad_and_vars[0][1]
grad_and_var = (grad, v)
average_grads.append(grad_and_var)
return average_grads
def train():
tf.logging.set_verbosity(tf.logging.DEBUG)
with tf.Graph().as_default():
with tf.name_scope('data_loader'):
with tf.device('/cpu:0'):
data_loader = data_batch.Data_Loader(dataset_dir=FLAGS.dataset_dir, split_sizes='train')
data_loader.get_dataset()
with tf.device('/cpu:0'):
global_step = tf.train.create_global_step()
boundaries = TRIAN_CONFIG['step_boundaries']
learning_rate = TRIAN_CONFIG['learning_rate']
tf.summary.scalar('learning_rate', learning_rate)
learning_rate = tf.train.piecewise_constant(global_step, boundaries, learning_rate )
optimizer = tf.train.GradientDescentOptimizer(learning_rate)
#add clones and calculate loss
tower_grads = []
summaries = []
with tf.variable_scope(tf.get_variable_scope()):
for i in range(0, FLAGS.gpu_num):
with tf.device('/gpu:%d' % i):
with tf.name_scope('%s_%d' % ('psenet_tower' , i)) as scope:
#dequeue
g_image, g_segmaps = data_loader.get_batch()
loss = tower_loss(scope, g_image, g_segmaps)
# Reuse variables for the next tower.
tf.get_variable_scope.reuse_variables()
# Retain the summaries from the final tower
summaries = tf.get_collection(tf.GraphKeys.SUMMARIES, scope)
grads = optimizer.compute_gradients(loss)
tower_grads.append(grads)
# We must calculate the mean of each gradient. Note that this is the
# synchronization point across all towers.
grads = average_gradients(tower_grads)
# Add a summary to track the learning rate.
summaries.append(tf.summary.scalar('learning_rate', learning_rate))
# Add histograms for gradients.
for grad, var in grads:
if grad is not None:
summaries.append(tf.summary.histogram(var.op.name + '/gradients', grad))
# Apply the gradients to adjust the shared variables
apply_gradient_op = optimizer.apply_gradients(grads, global_step=global_step)
# Track the moving averages of all trainable variables.
variable_averages = tf.train.ExponentialMovingAverage(TRIAN_CONFIG['MOVING_AVERAGE_DECAY'], global_step)
variable_averages_op = variable_averages.apply(tf.trainable_variables())
# Group all updates to into a single train op.
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies([variable_averages_op, update_ops]):
train_op = tf.group(apply_gradient_op, variable_averages)
saver = tf.train.Saver(tf.global_variables())
summary_writer = tf.summary.FileWriter(FLAGS.train_dir,
tf.get_default_graph())
init = tf.global_variables_initializer()
tf_config = tf.ConfigProto(allow_soft_placement=True)
with tf.Session(config=tf_config) as sess:
sess.run(init)
if FLAGS.checkpoint_path is not None:
print('from pevious checkpoint')
ckpt = tf.train.latest_checkpoint(FLAGS.train_dir)
saver.restore(sess, ckpt)
else:
if FLAGS.pretrained_model_path is not None:
variable_restore_op = slim.assign_from_checkpoint_fn(
FLAGS.checkpoint_path, slim.get_trainable_variables(),
ignore_missing_vars=FLAGS.ignore_missing_vars)
print(' pretrained model exists')
variable_restore_op(sess)
# start_step = global_step.eval() + 1
start_time = time.time()
if __name__ == '__main__':
train()