-
Notifications
You must be signed in to change notification settings - Fork 0
/
train_model.py
executable file
·159 lines (136 loc) · 6.77 KB
/
train_model.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
#!/usr/bin/python
#Heavily based on the translation example from google's tensorflow tutorials
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 numpy as np
import tensorflow as tf
#My stuff, and google's model
import utils
import seq2seq_model
tf.app.flags.DEFINE_float("learning_rate", 0.5, "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", 64,
"Batch size to use during training.")
tf.app.flags.DEFINE_integer("size", 1024, "Size of each model layer.")
tf.app.flags.DEFINE_integer("num_layers", 3, "Number of layers in the model.")
tf.app.flags.DEFINE_integer("en_vocab_size", 40000, "English vocabulary size.")
tf.app.flags.DEFINE_integer("fr_vocab_size", 40000, "French vocabulary size.")
tf.app.flags.DEFINE_string("data_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("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 = [(5, 10), (10, 15), (20, 25), (40, 50)]
def load_data_files(sourcePath, targetPath):
#A list of lists, each list is one bucket
data_set = [[] for _ in buckets]
with open(sourcePath, "r") as sourceFile, open(targetPath, "r") as targetFile:
source, target = sourceFile.readline(), targetFile.readline()
counter = 0
while source and target:
#Just to indicate something's happening
# counter += 1
# if counter % 1000 == 0:
# print "Reading line {0}".format(counter)
# sys.stdout.flush()
#Each line of a data file is a list of numbers with spaces between them.
sourceIDs = [int(x) for x in source.split()]
targetIDs = [int(x) for x in target.split()]
#TODO Should I append <<SEND>> symbols here?
#Put the data into the proper bucket
for bucket_id, (source_size, target_size) in enumerate(_buckets):
if len(sourceIDs) < source_size and len(targetIDs) < target_size:
data_set[bucket_id].append([sourceIDs, targetIDs])
break
source, target = sourceFile.readline(), targetFile.readline()
return data_set
def create_model(session, forward_only):
"""Create translation model and initialize or load parameters in session."""
model = seq2seq_model.Seq2SeqModel(
FLAGS.en_vocab_size, FLAGS.fr_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)
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(in_data_path, out_data_path):
#Prepare the data
with tf.Session() as sess:
model = create_model(sess, False)
#read the data into the buckets
#TODO seperate the training and the dev set, you're asking for overfitting here
dev_set = load_data_files(in_data_path, out_data_path)
train_set = load_data_files(in_data_path, out_data_path)
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, "translate.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 main(_):
train("stupid_deep_data/them_data.ids", "stupid_deep_data/me_data.ids")
if __name__ == "__main__":
tf.app.run()