-
Notifications
You must be signed in to change notification settings - Fork 14
/
transliterate.py
executable file
·347 lines (303 loc) · 15.1 KB
/
transliterate.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
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
"""Binary for training transliteration models and decoding from them.
Running this program without --decode will download the REV_brands corpus into
the directory specified as --data_dir and tokenize it in a very basic way,
and then start training a model saving checkpoints to --train_dir.
Running with --decode starts an interactive loop so you can see how
the current checkpoint transliterates English words into Hindi.
See the following papers for more information on neural translation models.
* http://arxiv.org/abs/1409.3215
* http://arxiv.org/abs/1409.0473
* http://arxiv.org/abs/1412.2007
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import math
import os
import random
import sys
import time
import codecs
import numpy as np
from six.moves import xrange # pylint: disable=redefined-builtin
import tensorflow as tf
import data_utils
import seq2seq_model
tf.app.flags.DEFINE_float("learning_rate", 0.001, "Learning rate.")
tf.app.flags.DEFINE_float("learning_rate_decay_factor", 0.99,
"Learning rate decays by this much.")
tf.app.flags.DEFINE_float("max_gradient_norm", 5.0,
"Clip gradients to this norm.")
tf.app.flags.DEFINE_integer("batch_size", 10,
"Batch size to use during training.")
tf.app.flags.DEFINE_integer("size", 256, "Size of each model layer.")
tf.app.flags.DEFINE_integer("num_layers", 2, "Number of layers in the model.")
tf.app.flags.DEFINE_integer("en_vocab_size", 40000, "English vocabulary size.")
tf.app.flags.DEFINE_integer("hn_vocab_size", 40000, "Hindi vocabulary size.")
tf.app.flags.DEFINE_string("data_dir", "/tmp", "Data directory")
tf.app.flags.DEFINE_string("transliterate_file_dir", "/tmp", "Data directory")
tf.app.flags.DEFINE_string("train_dir", "/tmp", "Training directory.")
tf.app.flags.DEFINE_integer("max_train_data_size", 0,
"Limit on the size of training data (0: no limit).")
tf.app.flags.DEFINE_integer("steps_per_checkpoint", 200,
"How many training steps to do per checkpoint.")
tf.app.flags.DEFINE_boolean("decode", False,
"Set to True for interactive decoding.")
tf.app.flags.DEFINE_boolean("transliterate_file", False,
"Set to True for evaluating.")
tf.app.flags.DEFINE_boolean("self_test", False,
"Run a self-test if this is set to True.")
FLAGS = tf.app.flags.FLAGS
# We use a number of buckets and pad to the closest one for efficiency.
# See seq2seq_model.Seq2SeqModel for details of how they work.
_buckets = [(3, 5), (5, 10), (10, 15), (15, 25),(25,35),(35,45)]
def read_data(source_path, target_path, max_size=None):
"""Read data from source and target files and put into buckets.
Args:
source_path: path to the files with token-ids for the source language.
target_path: path to the file with token-ids for the target language;
it must be aligned with the source file: n-th line contains the desired
output for n-th line from the source_path.
max_size: maximum number of lines to read, all other will be ignored;
if 0 or None, data files will be read completely (no limit).
Returns:
data_set: a list of length len(_buckets); data_set[n] contains a list of
(source, target) pairs read from the provided data files that fit
into the n-th bucket, i.e., such that len(source) < _buckets[n][0] and
len(target) < _buckets[n][1]; source and target are lists of token-ids.
"""
data_set = [[] for _ in _buckets]
with tf.gfile.GFile(source_path, mode="r") as source_file:
with tf.gfile.GFile(target_path, mode="r") as target_file:
source, target = source_file.readline(), target_file.readline()
counter = 0
while source and target and (not max_size or counter < max_size):
counter += 1
if counter % 10000 == 0:
print(" reading data line %d" % counter)
sys.stdout.flush()
source_ids = [int(x) for x in source.split()]
target_ids = [int(x) for x in target.split()]
target_ids.append(data_utils.EOS_ID)
for bucket_id, (source_size, target_size) in enumerate(_buckets):
if len(source_ids) < source_size and len(target_ids) < target_size:
data_set[bucket_id].append([source_ids, target_ids])
break
source, target = source_file.readline(), target_file.readline()
return data_set
def create_model(session, forward_only):
"""Create transliteration model and initialize or load parameters in session."""
model = seq2seq_model.Seq2SeqModel(
FLAGS.en_vocab_size, FLAGS.hn_vocab_size, _buckets,
FLAGS.size, FLAGS.num_layers, FLAGS.max_gradient_norm, FLAGS.batch_size,
FLAGS.learning_rate, FLAGS.learning_rate_decay_factor,
forward_only=forward_only,use_lstm=False)
ckpt = tf.train.get_checkpoint_state(FLAGS.train_dir)
if ckpt and tf.gfile.Exists(ckpt.model_checkpoint_path):
print("Reading model parameters from %s" % ckpt.model_checkpoint_path)
model.saver.restore(session, ckpt.model_checkpoint_path)
else:
print("Created model with fresh parameters.")
session.run(tf.initialize_all_variables())
return model
def train():
"""Train a en->hn transliteration model using REV_brandnames data."""
# Prepare REV_brandnames data.
print("Preparing REV data in %s" % FLAGS.data_dir)
en_train, hn_train, en_dev, hn_dev, _, _ = data_utils.prepare_rev_data(
FLAGS.data_dir, FLAGS.en_vocab_size, FLAGS.hn_vocab_size)
with tf.Session() as sess:
# Create model.
print("Creating %d layers of %d units." % (FLAGS.num_layers, FLAGS.size))
model = create_model(sess, False)
# Read data into buckets and compute their sizes.
print ("Reading development and training data (limit: %d)."
% FLAGS.max_train_data_size)
dev_set = read_data(en_dev, hn_dev)
train_set = read_data(en_train, hn_train, FLAGS.max_train_data_size)
train_bucket_sizes = [len(train_set[b]) for b in xrange(len(_buckets))]
train_total_size = float(sum(train_bucket_sizes))
# A bucket scale is a list of increasing numbers from 0 to 1 that we'll use
# to select a bucket. Length of [scale[i], scale[i+1]] is proportional to
# the size if i-th training bucket, as used later.
train_buckets_scale = [sum(train_bucket_sizes[:i + 1]) / train_total_size
for i in xrange(len(train_bucket_sizes))]
# This is the training loop.
step_time, loss = 0.0, 0.0
current_step = 0
previous_losses = []
while True:
# Choose a bucket according to data distribution. We pick a random number
# in [0, 1] and use the corresponding interval in train_buckets_scale.
random_number_01 = np.random.random_sample()
bucket_id = min([i for i in xrange(len(train_buckets_scale))
if train_buckets_scale[i] > random_number_01])
# Get a batch and make a step.
start_time = time.time()
encoder_inputs, decoder_inputs, target_weights = model.get_batch(
train_set, bucket_id)
_, step_loss, _ = model.step(sess, encoder_inputs, decoder_inputs,
target_weights, bucket_id, False)
step_time += (time.time() - start_time) / FLAGS.steps_per_checkpoint
loss += step_loss / FLAGS.steps_per_checkpoint
current_step += 1
# Once in a while, we save checkpoint, print statistics, and run evals.
if current_step % FLAGS.steps_per_checkpoint == 0:
# Print statistics for the previous epoch.
perplexity = math.exp(loss) if loss < 300 else float('inf')
print ("global step %d learning rate %.4f step-time %.2f perplexity "
"%.2f" % (model.global_step.eval(), model.learning_rate.eval(),
step_time, perplexity))
# Decrease learning rate if no improvement was seen over last 3 times.
if len(previous_losses) > 2 and loss > max(previous_losses[-3:]):
sess.run(model.learning_rate_decay_op)
previous_losses.append(loss)
# Save checkpoint and zero timer and loss.
checkpoint_path = os.path.join(FLAGS.train_dir, "transliterate.ckpt")
model.saver.save(sess, checkpoint_path, global_step=model.global_step)
step_time, loss = 0.0, 0.0
# Run evals on development set and print their perplexity.
for bucket_id in xrange(len(_buckets)):
if len(dev_set[bucket_id]) == 0:
print(" eval: empty bucket %d" % (bucket_id))
continue
encoder_inputs, decoder_inputs, target_weights = model.get_batch(
dev_set, bucket_id)
_, eval_loss, _ = model.step(sess, encoder_inputs, decoder_inputs,
target_weights, bucket_id, True)
eval_ppx = math.exp(eval_loss) if eval_loss < 300 else float('inf')
print(" eval: bucket %d perplexity %.2f" % (bucket_id, eval_ppx))
sys.stdout.flush()
def decode():
with tf.Session() as sess:
# Create model and load parameters.
model = create_model(sess, True)
model.batch_size = 1 # We decode one word at a time.
# Load vocabularies.
en_vocab_path = os.path.join(FLAGS.data_dir,
"vocab%d.en" % FLAGS.en_vocab_size)
hn_vocab_path = os.path.join(FLAGS.data_dir,
"vocab%d.hn" % FLAGS.hn_vocab_size)
en_vocab, _ = data_utils.initialize_vocabulary(en_vocab_path)
_, rev_hn_vocab = data_utils.initialize_vocabulary(hn_vocab_path)
# Decode from standard input.
sys.stdout.write("> ")
sys.stdout.flush()
word = sys.stdin.readline()
while word:
word = word.lower()
char_list_new = list(word)
word = " ".join(char_list_new)
# Get token-ids for the input word.
token_ids = data_utils.word_to_token_ids(tf.compat.as_bytes(word), en_vocab)
# Which bucket does it belong to?
bucket_id = min([b for b in xrange(len(_buckets))
if _buckets[b][0] > len(token_ids)])
# Get a 1-element batch to feed the word to the model.
encoder_inputs, decoder_inputs, target_weights = model.get_batch(
{bucket_id: [(token_ids, [])]}, bucket_id)
# Get output logits for the word.
_, _, output_logits = model.step(sess, encoder_inputs, decoder_inputs,
target_weights, bucket_id, True)
# This is a greedy decoder - outputs are just argmaxes of output_logits.
outputs = [int(np.argmax(logit, axis=1)) for logit in output_logits]
# If there is an EOS symbol in outputs, cut them at that point.
if data_utils.EOS_ID in outputs:
outputs = outputs[:outputs.index(data_utils.EOS_ID)]
# Print out Hindi word corresponding to outputs.
print("".join([tf.compat.as_str(rev_hn_vocab[output]) for output in outputs]))
print("> ", end="")
sys.stdout.flush()
word = sys.stdin.readline()
def self_test():
"""Test the translation model."""
with tf.Session() as sess:
print("Self-test for neural transliteration model.")
# Create model with vocabularies of 10, 2 small buckets, 2 layers of 32.
model = seq2seq_model.Seq2SeqModel(10, 10, [(3, 3), (6, 6)], 32, 2,
5.0, 32, 0.3, 0.99, num_samples=8)
sess.run(tf.initialize_all_variables())
# Fake data set for both the (3, 3) and (6, 6) bucket.
data_set = ([([1, 1], [2, 2]), ([3, 3], [4]), ([5], [6])],
[([1, 1, 1, 1, 1], [2, 2, 2, 2, 2]), ([3, 3, 3], [5, 6])])
for _ in xrange(5): # Train the fake model for 5 steps.
bucket_id = random.choice([0, 1])
encoder_inputs, decoder_inputs, target_weights = model.get_batch(
data_set, bucket_id)
model.step(sess, encoder_inputs, decoder_inputs, target_weights,
bucket_id, False)
def evaluate():
"""Generate an evaluation output of the model.
Takes the directory path for evaluation from FLAGS and writes an output file to the same directory"""
with tf.Session() as sess:
# Create model and load parameters.
model = create_model(sess, True)
model.batch_size = 1 # We decode one word at a time.
# Load vocabularies.
en_vocab_path = os.path.join(FLAGS.data_dir,
"vocab%d.en" % FLAGS.en_vocab_size)
hn_vocab_path = os.path.join(FLAGS.data_dir,
"vocab%d.hn" % FLAGS.hn_vocab_size)
en_vocab, _ = data_utils.initialize_vocabulary(en_vocab_path)
_, rev_hn_vocab = data_utils.initialize_vocabulary(hn_vocab_path)
#path for loading the evaluation file
en_eval_path = os.path.join(FLAGS.transliterate_file_dir,'test.en')
print('Transliterating '+en_eval_path)
#Path to save the output file
result_path = os.path.join(FLAGS.transliterate_file_dir,'result.txt')
print('Results will be stored in '+result_path)
en_eval_list = []
file_content_output = []
print('reading input file')
with open(en_eval_path) as fp:
for line in fp:
char_list = list(line)
space_separated = ' '.join(char_list)
en_eval_list.append(space_separated)
print('decoding input file')
for i,word in enumerate(en_eval_list):
word = word.lower()
char_list_new = list(word)
word = " ".join(char_list_new)
# Get token-ids for the input word.
token_ids = data_utils.word_to_token_ids(tf.compat.as_bytes(word), en_vocab)
# Which bucket does it belong to?
bucket_list = [b for b in xrange(len(_buckets))
if _buckets[b][0] > len(token_ids)]
#bucket_id = min([b for b in xrange(len(_buckets))
# if _buckets[b][0] > len(token_ids)])
if len(bucket_list) == 0:
print('could not find bucket')
continue
bucket_id = min(bucket_list)
# Get a 1-element batch to feed the word to the model.
encoder_inputs, decoder_inputs, target_weights = model.get_batch(
{bucket_id: [(token_ids, [])]}, bucket_id)
# Get output logits for the word.
_, _, output_logits = model.step(sess, encoder_inputs, decoder_inputs,
target_weights, bucket_id, True)
# This is a greedy decoder - outputs are just argmaxes of output_logits.
outputs = [int(np.argmax(logit, axis=1)) for logit in output_logits]
# If there is an EOS symbol in outputs, cut them at that point.
if data_utils.EOS_ID in outputs:
outputs = outputs[:outputs.index(data_utils.EOS_ID)]
# Print out Hindi word corresponding to outputs.
hn_output = "".join([tf.compat.as_str(rev_hn_vocab[output]) for output in outputs])
if i%100 == 0:
print(str(i)+' out of ' + str(len(en_eval_list)) +' words decoded\n English Input: ' + word + '\t Hindi Output: ' + hn_output)
file_content_output.append([word,hn_output])
print('done generating the output file!!!')
fc_str = '\n'.join(['\t'.join(row) for row in file_content_output])
f = codecs.open(result_path, encoding='utf-8', mode='wb')
f.write(fc_str.decode('utf-8'))
def main(_):
if FLAGS.self_test:
self_test()
elif FLAGS.decode:
decode()
elif FLAGS.transliterate_file:
evaluate()
else:
train()
if __name__ == "__main__":
tf.app.run()