-
Notifications
You must be signed in to change notification settings - Fork 122
/
data_load.py
218 lines (183 loc) · 7.7 KB
/
data_load.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
# -*- coding: utf-8 -*-
#/usr/bin/python2
from functools import wraps
import threading
from tensorflow.python.platform import tf_logging as logging
from params import Params
import numpy as np
import tensorflow as tf
from process import *
from sklearn.model_selection import train_test_split
# Adapted from the `sugartensor` code.
# https://github.com/buriburisuri/sugartensor/blob/master/sugartensor/sg_queue.py
def producer_func(func):
r"""Decorates a function `func` as producer_func.
Args:
func: A function to decorate.
"""
@wraps(func)
def wrapper(inputs, dtypes, capacity, num_threads):
r"""
Args:
inputs: A inputs queue list to enqueue
dtypes: Data types of each tensor
capacity: Queue capacity. Default is 32.
num_threads: Number of threads. Default is 1.
"""
# enqueue function
def enqueue_func(sess, op):
# read data from source queue
data = func(sess.run(inputs))
# create feeder dict
feed_dict = {}
for ph, col in zip(placeholders, data):
feed_dict[ph] = col
# run session
sess.run(op, feed_dict=feed_dict)
# create place holder list
placeholders = []
for dtype in dtypes:
placeholders.append(tf.placeholder(dtype=dtype))
# create FIFO queue
queue = tf.FIFOQueue(capacity, dtypes=dtypes)
# enqueue operation
enqueue_op = queue.enqueue(placeholders)
# create queue runner
runner = _FuncQueueRunner(enqueue_func, queue, [enqueue_op] * num_threads)
# register to global collection
tf.train.add_queue_runner(runner)
# return de-queue operation
return queue.dequeue()
return wrapper
class _FuncQueueRunner(tf.train.QueueRunner):
def __init__(self, func, queue=None, enqueue_ops=None, close_op=None,
cancel_op=None, queue_closed_exception_types=None,
queue_runner_def=None):
# save ad-hoc function
self.func = func
# call super()
super(_FuncQueueRunner, self).__init__(queue, enqueue_ops, close_op, cancel_op,
queue_closed_exception_types, queue_runner_def)
# pylint: disable=broad-except
def _run(self, sess, enqueue_op, coord=None):
if coord:
coord.register_thread(threading.current_thread())
decremented = False
try:
while True:
if coord and coord.should_stop():
break
try:
self.func(sess, enqueue_op) # call enqueue function
except self._queue_closed_exception_types: # pylint: disable=catching-non-exception
# This exception indicates that a queue was closed.
with self._lock:
self._runs_per_session[sess] -= 1
decremented = True
if self._runs_per_session[sess] == 0:
try:
sess.run(self._close_op)
except Exception as e:
# Intentionally ignore errors from close_op.
logging.vlog(1, "Ignored exception: %s", str(e))
return
except Exception as e:
# This catches all other exceptions.
if coord:
coord.request_stop(e)
else:
logging.error("Exception in QueueRunner: %s", str(e))
with self._lock:
self._exceptions_raised.append(e)
raise
finally:
# Make sure we account for all terminations: normal or errors.
if not decremented:
with self._lock:
self._runs_per_session[sess] -= 1
def load_data(dir_):
# Target indices
indices = load_target(dir_ + Params.target_dir)
# Load question data
print("Loading question data...")
q_word_ids, _ = load_word(dir_ + Params.q_word_dir)
q_char_ids, q_char_len, q_word_len = load_char(dir_ + Params.q_chars_dir)
# Load passage data
print("Loading passage data...")
p_word_ids, _ = load_word(dir_ + Params.p_word_dir)
p_char_ids, p_char_len, p_word_len = load_char(dir_ + Params.p_chars_dir)
# Get max length to pad
p_max_word = Params.max_p_len#np.max(p_word_len)
p_max_char = Params.max_char_len#,max_value(p_char_len))
q_max_word = Params.max_q_len#,np.max(q_word_len)
q_max_char = Params.max_char_len#,max_value(q_char_len))
# pad_data
print("Preparing data...")
p_word_ids = pad_data(p_word_ids,p_max_word)
q_word_ids = pad_data(q_word_ids,q_max_word)
p_char_ids = pad_char_data(p_char_ids,p_max_char,p_max_word)
q_char_ids = pad_char_data(q_char_ids,q_max_char,q_max_word)
# to numpy
indices = np.reshape(np.asarray(indices,np.int32),(-1,2))
p_word_len = np.reshape(np.asarray(p_word_len,np.int32),(-1,1))
q_word_len = np.reshape(np.asarray(q_word_len,np.int32),(-1,1))
# p_char_len = pad_data(p_char_len,p_max_word)
# q_char_len = pad_data(q_char_len,q_max_word)
p_char_len = pad_char_len(p_char_len, p_max_word, p_max_char)
q_char_len = pad_char_len(q_char_len, q_max_word, q_max_char)
for i in range(p_word_len.shape[0]):
if p_word_len[i,0] > p_max_word:
p_word_len[i,0] = p_max_word
for i in range(q_word_len.shape[0]):
if q_word_len[i,0] > q_max_word:
q_word_len[i,0] = q_max_word
# shapes of each data
shapes=[(p_max_word,),(q_max_word,),
(p_max_word,p_max_char,),(q_max_word,q_max_char,),
(1,),(1,),
(p_max_word,),(q_max_word,),
(2,)]
return ([p_word_ids, q_word_ids,
p_char_ids, q_char_ids,
p_word_len, q_word_len,
p_char_len, q_char_len,
indices], shapes)
def get_dev():
devset, shapes = load_data(Params.dev_dir)
indices = devset[-1]
# devset = [np.reshape(input_, shapes[i]) for i,input_ in enumerate(devset)]
dev_ind = np.arange(indices.shape[0],dtype = np.int32)
np.random.shuffle(dev_ind)
return devset, dev_ind
def get_batch(is_training = True):
"""Loads training data and put them in queues"""
with tf.device('/cpu:0'):
# Load dataset
input_list, shapes = load_data(Params.train_dir if is_training else Params.dev_dir)
indices = input_list[-1]
train_ind = np.arange(indices.shape[0],dtype = np.int32)
np.random.shuffle(train_ind)
size = Params.data_size
if Params.data_size > indices.shape[0] or Params.data_size == -1:
size = indices.shape[0]
ind_list = tf.convert_to_tensor(train_ind[:size])
# Create Queues
ind_list = tf.train.slice_input_producer([ind_list], shuffle=True)
@producer_func
def get_data(ind):
'''From `_inputs`, which has been fetched from slice queues,
then enqueue them again.
'''
return [np.reshape(input_[ind], shapes[i]) for i,input_ in enumerate(input_list)]
data = get_data(inputs=ind_list,
dtypes=[np.int32]*9,
capacity=Params.batch_size*32,
num_threads=6)
# create batch queues
batch = tf.train.batch(data,
shapes=shapes,
num_threads=2,
batch_size=Params.batch_size,
capacity=Params.batch_size*32,
dynamic_pad=True)
return batch, size // Params.batch_size