-
Notifications
You must be signed in to change notification settings - Fork 53
/
chiron_rcnn_train.py
227 lines (222 loc) · 11.7 KB
/
chiron_rcnn_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
#!/usr/bin/env python2
# -*- coding: utf-8 -*-
# Copyright 2017 The Chiron Authors. All Rights Reserved.
#
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at http://mozilla.org/MPL/2.0/.
#
# Created on Mon Mar 27 14:04:57 2017
# from rnn import rnn_layers
from __future__ import absolute_import
from __future__ import print_function
import os
import sys
import time
import argparse
import tensorflow as tf
import chiron.chiron_model as model
from chiron.chiron_input import read_raw_data_sets
from chiron.chiron_input import read_cache_dataset
from tensorflow.python.ops import variables
from six.moves import range
DEFAULT_OFFSET = 10
tf.logging.set_verbosity(tf.logging.ERROR)
def save_hyper_parameter():
"""
TODO: Function to save the hyper parameter.
"""
def compile_train_graph(config,hp):
class net:
pass
net.training = tf.placeholder(tf.bool)
net.global_step = tf.get_variable('global_step', trainable=False, shape=(),
dtype=tf.int32,
initializer=tf.zeros_initializer())
net.x = tf.placeholder(tf.float32, shape=[hp.batch_size, hp.sequence_len])
net.seq_length = tf.placeholder(tf.int32, shape=[hp.batch_size])
net.y_indexs = tf.placeholder(tf.int64)
net.y_values = tf.placeholder(tf.int32)
net.y_shape = tf.placeholder(tf.int64)
net.y = tf.SparseTensor(net.y_indexs, net.y_values, net.y_shape)
net.logits, net.ratio = model.inference(net.x, net.seq_length, net.training,hp.sequence_len,configure = config)
if 'fl_gamma' in config.keys():
net.ctc_loss = model.loss(net.logits, net.seq_length, net.y, fl_gamma = config['fl_gamma'])
else:
net.ctc_loss = model.loss(net.logits, net.seq_length, net.y)
net.opt = model.train_opt(hp.step_rate,
hp.max_steps,
global_step=net.global_step,
opt_name = config['opt_method'])
if hp.gradient_clip is None:
net.step = net.opt.minimize(net.ctc_loss,global_step = net.global_step)
else:
net.gradients, net.variables = zip(*net.opt.compute_gradients(net.ctc_loss))
net.gradients = [None if gradient is None else tf.clip_by_norm(gradient, hp.gradient_clip) for gradient in net.gradients]
net.step = net.opt.apply_gradients(zip(net.gradients, net.variables),global_step = net.global_step)
net.error,net.errors,net.y_ = model.prediction(net.logits, net.seq_length, net.y)
net.init = tf.global_variables_initializer()
net.variable_to_restore=set(variables._all_saveable_objects()+tf.moving_average_variables())
net.saver = tf.train.Saver(var_list=net.variable_to_restore,
save_relative_paths=True)
net.summary = tf.summary.merge_all()
return net
def train():
default_config = os.path.join(FLAGS.log_dir,FLAGS.model_name,'model.json')
if FLAGS.retrain:
if os.path.isfile(default_config):
config_file = default_config
else:
raise ValueError("Model Json file has not been found in model log directory")
else:
config_file = FLAGS.configure
config = model.read_config(config_file)
print("Begin training using following setting:")
with open(os.path.join(FLAGS.log_dir,FLAGS.model_name,'train_config'),'w+') as log_f:
for pro in dir(FLAGS):
if not pro.startswith('_'):
print("%s:%s"%(pro,getattr(FLAGS,pro)))
log_f.write("%s:%s\n"%(pro,getattr(FLAGS,pro)))
net = compile_train_graph(config,FLAGS)
sess = tf.Session(config=tf.ConfigProto(inter_op_parallelism_threads=FLAGS.threads,
intra_op_parallelism_threads=FLAGS.threads,
allow_soft_placement=True))
if FLAGS.retrain == False:
sess.run(net.init)
print("Model init finished, begin loading data. \n")
else:
net.saver.restore(sess, tf.train.latest_checkpoint(
FLAGS.log_dir + FLAGS.model_name))
print("Model loaded finished, begin loading data. \n")
summary_writer = tf.summary.FileWriter(
FLAGS.log_dir + FLAGS.model_name + '/summary/', sess.graph)
model.save_model(default_config,config)
train_ds,valid_ds = generate_train_valid_datasets(initial_offset = DEFAULT_OFFSET)
start = time.time()
resample_n = 0
for i in range(FLAGS.max_steps):
if FLAGS.resample_after_epoch == 0:
pass
elif train_ds.epochs_completed >= FLAGS.resample_after_epoch:
train_ds,valid_ds = generate_train_valid_datasets(initial_offset = resample_n*FLAGS.offset_increment + DEFAULT_OFFSET)
batch_x, seq_len, batch_y = train_ds.next_batch(FLAGS.batch_size)
indxs, values, shape = batch_y
feed_dict = {net.x: batch_x, net.seq_length: seq_len / net.ratio, net.y_indexs: indxs,
net.y_values: values, net.y_shape: shape,
net.training: True}
loss_val, _ = sess.run([net.ctc_loss, net.step], feed_dict=feed_dict)
if i % 10 == 0:
global_step_val = tf.train.global_step(sess, net.global_step)
valid_x, valid_len, valid_y = valid_ds.next_batch(FLAGS.batch_size)
indxs, values, shape = valid_y
feed_dict = {net.x: valid_x, net.seq_length: valid_len / net.ratio,
net.y_indexs: indxs, net.y_values: values, net.y_shape: shape,
net.training: True}
error_val = sess.run(net.error, feed_dict=feed_dict)
# x_val,errors_val,y_predict,y = sess.run([x,errors,y_,y],feed_dict = feed_dict)
# predict_seq,_ = sparse2dense([y_predict,0])
# true_seq,_ = sparse2dense([[y],0])
end = time.time()
print(
"Step %d/%d Epoch %d, batch number %d, train_loss: %5.3f validate_edit_distance: %5.3f Elapsed Time/step: %5.3f" \
% (i, FLAGS.max_steps, train_ds.epochs_completed,
train_ds.index_in_epoch, loss_val, error_val,
(end - start) / (i + 1)))
net.saver.save(sess, FLAGS.log_dir + FLAGS.model_name + '/model.ckpt',
global_step=global_step_val)
summary_str = sess.run(net.summary, feed_dict=feed_dict)
summary_writer.add_summary(summary_str, global_step=global_step_val)
summary_writer.flush()
global_step_val = tf.train.global_step(sess, net.global_step)
print("Model %s saved." % (FLAGS.log_dir + FLAGS.model_name))
print("Reads number %d" % (train_ds.reads_n))
net.saver.save(sess, FLAGS.log_dir + FLAGS.model_name + '/final.ckpt',
global_step=global_step_val)
def generate_train_valid_datasets(initial_offset = 10):
if FLAGS.read_cache:
train_ds = read_cache_dataset(FLAGS.train_cache)
if FLAGS.validation is not None:
valid_ds = read_cache_dataset(FLAGS.valid_cache)
else:
valid_ds = train_ds
if train_ds.event.shape[1]!=FLAGS.sequence_len:
raise ValueError("The event length of training cached dataset %d is inconsistent with given sequene_len %d"%(train_ds.event.shape()[1],FLAGS.sequence_len))
if valid_ds.event.shape[1]!=FLAGS.sequence_len:
raise ValueError("The event length of training cached dataset %d is inconsistent with given sequene_len %d"%(valid_ds.event.shape()[1],FLAGS.sequence_len))
return train_ds,valid_ds
sys.stdout.write("Begin reading training dataset.\n")
train_ds = read_raw_data_sets(FLAGS.data_dir,
FLAGS.train_cache,
FLAGS.sequence_len,
k_mer=FLAGS.k_mer,
max_segments_num=FLAGS.segments_num,
skip_start = initial_offset)
sys.stdout.write("Begin reading validation dataset.\n")
if FLAGS.validation is not None:
valid_ds = read_raw_data_sets(FLAGS.validation,
FLAGS.valid_cache,
FLAGS.sequence_len,
k_mer=FLAGS.k_mer,
max_segments_num=FLAGS.segments_num)
else:
valid_ds = train_ds
return train_ds,valid_ds
def run(args):
global FLAGS
FLAGS = args
if FLAGS.train_cache is None:
FLAGS.train_cache = os.path.join(FLAGS.log_dir,FLAGS.model_name) + '/train_cache.hdf5'
if (FLAGS.valid_cache is None) and (FLAGS.validation is not None):
FLAGS.valid_cache = os.path.join(FLAGS.log_dir,FLAGS.model_name) + '/valid_cache.hdf5'
FLAGS.data_dir = FLAGS.data_dir + os.path.sep
FLAGS.log_dir = FLAGS.log_dir + os.path.sep
train()
if __name__ == "__main__":
parser = argparse.ArgumentParser(
description='Training model with tfrecord file')
parser.add_argument('-i', '--data_dir', required = True,
help="Directory that store the tfrecord files.")
parser.add_argument('-o', '--log_dir', required = True ,
help="log directory that store the training model.")
parser.add_argument('-m', '--model_name', required = True,
help='model_name')
parser.add_argument('-v', '--validation', default = None,
help="validation data folder, default is None, which use the train dataset.")
parser.add_argument('--train_cache', default=None, help="Cache file for training dataset.")
parser.add_argument('--valid_cache', default=None, help="Cache file for validation dataset.")
parser.add_argument('-s', '--sequence_len', type=int, default=400,
help='the length of sequence')
parser.add_argument('-b', '--batch_size', type=int, default=300,
help='Batch size')
parser.add_argument('-t', '--step_rate', type=float, default=4e-3,
help='Step rate')
parser.add_argument('-x', '--max_steps', type=int, default=10000,
help='Maximum step')
parser.add_argument('-n', '--segments_num', type = int, default = None,
help='Maximum number of segments read into the training queue, default(None) read all segments.')
parser.add_argument('--configure', default = None,
help="Model structure configure json file.")
parser.add_argument('-k', '--k_mer', default=1, help='Output k-mer size')
parser.add_argument('--resample_after_epoch',
type = int,
default = 0,
help='Resample the reads data every n epoches, with an increasing initial offset.')
parser.add_argument('--threads',
type = int,
default = 0,
help='The threads that available, if 0 use all threads that can be found.')
parser.add_argument('--offset_increment',
type = int,
default = 3,
help='The increament of initial offset if the resample_after_epoch has been set.')
parser.add_argument('--gradient_clip',
type = float,
default = None,
help = 'Clip the gradient by the gradient_clip x normalization, a good estimate is 5.')
parser.add_argument('--retrain', dest='retrain', action='store_true',
help='Set retrain to true')
parser.add_argument('--read_cache',dest='read_cache',action='store_true',
help="Read from cached hdf5 file.")
parser.set_defaults(retrain=False)
args = parser.parse_args(sys.argv[1:])
run(args)