-
Notifications
You must be signed in to change notification settings - Fork 15
/
train.py
194 lines (157 loc) · 6.9 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
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
# -*- coding: utf-8 -*-
# @Author : Junru_Lu
# @File : train.py
# @Software: PyCharm
# @Environment : Python 3.6+
# @Reference1 : https://zhuanlan.zhihu.com/p/31638132
# @Reference2 : https://github.com/likejazz/Siamese-LSTM
# 基础包
from time import time
import pandas as pd
from sklearn.model_selection import train_test_split
import keras
from gensim.models import KeyedVectors
from keras.models import Model
from keras.layers import Input, Embedding, LSTM, Dense, Flatten, Activation, RepeatVector, Permute, Lambda, \
Bidirectional, TimeDistributed, Dropout, Conv1D, GlobalMaxPool1D
from keras.layers.merge import multiply, concatenate
import keras.backend as K
from util import make_w2v_embeddings, split_and_zero_padding, ManDist
'''
本配置文件用于训练孪生网络
'''
# ------------------预加载------------------ #
# 中英文训练选择，默认使用英文训练集
s = input("type cn or en:")
if s == 'cn':
TRAIN_CSV = './data/quora_train_segmented.csv'
flag = 'cn'
embedding_path = 'CnCorpus-vectors-negative64.bin'
embedding_dim = 64
max_seq_length = 20
savepath = './data/cn_SiameseLSTM.h5'
else:
TRAIN_CSV = './data/quora_train.csv'
flag = 'en'
embedding_path = 'GoogleNews-vectors-negative300.bin'
embedding_dim = 300
max_seq_length = 10
savepath = './data/en_SiameseLSTM.h5'
# 是否启用预训练的词向量，默认使用随机初始化的词向量
o = input("type yes or no for choosing pre-trained w2v or not:")
if o == 'yes':
# 加载词向量
print("Loading word2vec model(it may takes 2-3 mins) ...")
embedding_dict = KeyedVectors.load_word2vec_format(embedding_path, binary=True)
else:
embedding_dict = {}
# 读取并加载训练集
train_df = pd.read_csv(TRAIN_CSV)
for q in ['question1', 'question2']:
train_df[q + '_n'] = train_df[q]
# 将训练集词向量化
train_df, embeddings = make_w2v_embeddings(flag, embedding_dict, train_df, embedding_dim=embedding_dim)
'''
把训练数据从：
question1 question2 is_duplicate
借 呗 还款 信息 借 呗 还款 日期 0
变成：
question1 question2 is_duplicate question1_n question2_n
借 呗 还款 信息 借 呗 还款 日期 0 借 呗 还款 信息 借 呗 还款 日期
变成id以后：
question1 question2 is_duplicate question1_n question2_n
借 呗 还款 信息 借 呗 还款 日期 0 [31, 639] [31, 255]
'''
# 分割训练集
X = train_df[['question1_n', 'question2_n']]
Y = train_df['is_duplicate']
X_train, X_validation, Y_train, Y_validation = train_test_split(X, Y, test_size=0.1)
X_train = split_and_zero_padding(X_train, max_seq_length)
X_validation = split_and_zero_padding(X_validation, max_seq_length)
# 将标签转化为数值
Y_train = Y_train.values
Y_validation = Y_validation.values
# 确认数据准备完毕且正确
assert X_train['left'].shape == X_train['right'].shape
assert len(X_train['left']) == len(Y_train)
# -----------------基础函数------------------ #
def shared_model(_input):
# 词向量化
embedded = Embedding(len(embeddings), embedding_dim, weights=[embeddings], input_shape=(max_seq_length,),
trainable=False)(_input)
# 多层Bi-LSTM
activations = Bidirectional(LSTM(n_hidden, return_sequences=True), merge_mode='concat')(embedded)
activations = Bidirectional(LSTM(n_hidden, return_sequences=True), merge_mode='concat')(activations)
# dropout
# activations = Dropout(0.5)(activations)
# Attention
attention = TimeDistributed(Dense(1, activation='tanh'))(activations)
attention = Flatten()(attention)
attention = Activation('softmax')(attention)
attention = RepeatVector(n_hidden * 2)(attention)
attention = Permute([2, 1])(attention)
sent_representation = multiply([activations, attention])
sent_representation = Lambda(lambda xin: K.sum(xin, axis=1))(sent_representation)
# dropout
# sent_representation = Dropout(0.1)(sent_representation)
return sent_representation
def shared_model_cnn(_input):
# 词向量化
embedded = Embedding(len(embeddings), embedding_dim, weights=[embeddings], input_shape=(max_seq_length,),
trainable=False)(_input)
# CNN
activations = Conv1D(250, kernel_size=5, activation='relu')(embedded)
activations = GlobalMaxPool1D()(activations)
activations = Dense(250, activation='relu')(activations)
activations = Dropout(0.3)(activations)
activations = Dense(1, activation='sigmoid')(activations)
return activations
# -----------------主函数----------------- #
if __name__ == '__main__':
# 超参
batch_size = 1024
n_epoch = 9
n_hidden = 50
left_input = Input(shape=(max_seq_length,), dtype='float32')
right_input = Input(shape=(max_seq_length,), dtype='float32')
left_sen_representation = shared_model(left_input)
right_sen_representation = shared_model(right_input)
# 引入曼哈顿距离，把得到的变换concat上原始的向量再通过一个多层的DNN做了下非线性变换、sigmoid得相似度
# 没有使用https://zhuanlan.zhihu.com/p/31638132中提到的马氏距离，尝试了曼哈顿距离、点乘和cos，效果曼哈顿最好
man_distance = ManDist()([left_sen_representation, right_sen_representation])
sen_representation = concatenate([left_sen_representation, right_sen_representation, man_distance])
similarity = Dense(1, activation='sigmoid')(Dense(2)(Dense(4)(Dense(16)(sen_representation))))
model = Model(inputs=[left_input, right_input], outputs=[similarity])
model.compile(loss='mean_squared_error', optimizer=keras.optimizers.Adam(), metrics=['accuracy'])
model.summary()
training_start_time = time()
malstm_trained = model.fit([X_train['left'], X_train['right']], Y_train,
batch_size=batch_size, epochs=n_epoch,
validation_data=([X_validation['left'], X_validation['right']], Y_validation))
training_end_time = time()
print("Training time finished.\n%d epochs in %12.2f" % (n_epoch, training_end_time - training_start_time))
# Plot accuracy
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
plt.subplot(211)
plt.plot(malstm_trained.history['acc'])
plt.plot(malstm_trained.history['val_acc'])
plt.title('Model Accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
# Plot loss
plt.subplot(212)
plt.plot(malstm_trained.history['loss'])
plt.plot(malstm_trained.history['val_loss'])
plt.title('Model Loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper right')
plt.tight_layout(h_pad=1.0)
plt.savefig('./data/history-graph.png')
model.save(savepath)
print(str(malstm_trained.history['val_acc'][-1])[:6] +
"(max: " + str(max(malstm_trained.history['val_acc']))[:6] + ")")
print("Done.")