-
Notifications
You must be signed in to change notification settings - Fork 1
/
train.py
84 lines (67 loc) · 2.38 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
import tensorflow as tf
import numpy as np
import codecs
from gensim.models.word2vec import Word2Vec
from stk import *
from configs import *
from preprocessor import *
flags = tf.flags
flags.DEFINE_string(
'model',
'test', 'A type of model. Possible options are: large, test.')
flags.DEFINE_string(
'data_path',
'./data/simple/simple.rd.tk', 'The path to correctly-formatted data.')
flags.DEFINE_boolean(
'tensorboard',
False, 'Whether to write data to a TensorBoard summary.')
flags.DEFINE_integer(
'vocab_size',
8000, 'Vocaburaly size.')
FLAGS = flags.FLAGS
def get_config():
if FLAGS.model == 'small':
return SmallConfig()
elif FLAGS.model == 'test':
return TestConfig()
else:
return ValueError('Invalid model: ', FLAGS.model)
if __name__ == '__main__':
if not FLAGS.data_path:
raise ValueError('Must set --data_path.')
config = get_config()
if FLAGS.vocab_size:
config.vocab_size = FLAGS.vocab_size
train_file_path = './data/result/train.kr'
test_file_path = './data/result/test.kr'
data_file = codecs.open(FLAGS.data_path, 'r', 'utf-8')
data = data_file.readlines()
data_file.close()
sent_list = [d.split() for d in data]
word2idx, idx2word = build_dict(sent_list, config)
model = Word2Vec.load('./tmp/word2vec.model')
embeddings = np.zeros((config.vocab_size, config.embedding_size))
for (w, i) in word2idx.items():
try:
embeddings[i] = model.wv[w]
except:
pass
x_data, y_data = make_parallel_data(sent_list, word2idx)
x_divide, y_divide = divide_sentences(x_data, y_data, config.seq_size)
train_file = codecs.open(train_file_path, 'w', 'utf-8')
test_file = codecs.open(test_file_path, 'w', 'utf-8')
num_data = len(x_divide)
if num_data < 5000:
num_test = int(num_data / 5)
else:
num_test = 1000
num_train = num_data - num_test
for i in range(num_test):
test_file.write(' '.join(x_divide[i])+'\n'+' '.join(y_divide[i])+'\n')
for i in range(num_test, num_data):
train_file.write(' '.join(x_divide[i])+'\n'+' '.join(y_divide[i])+'\n')
train_file.close()
test_file.close()
stk = SentenceTokenizer(config=config, embeddings=embeddings)
train_files = [train_file_path]
stk.train(train_files, config.epochs, num_train)