/
train_textcnn_keras.py
129 lines (112 loc) · 5.32 KB
/
train_textcnn_keras.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
import numpy as np
import pandas as pd
from keras.models import Model
from keras.layers import Dense, Embedding, Input, Dropout
from keras.layers.merge import concatenate
from keras.layers import Conv1D, BatchNormalization, Activation, GlobalMaxPooling1D
from data_utils.evaluation import *
from sklearn.metrics import f1_score
from keras.utils.vis_utils import plot_model
embedding_dim = 256 # embedding layer dimension is fixed to 256
num_words = 40000
maxlen = 400
label_type = 'accusation'
##################################################
# data and label
train_fact_pad_seq = np.load('./variables/pad_sequences/train_pad_%d_%d.npy' % (maxlen, num_words))
valid_fact_pad_seq = np.load('./variables/pad_sequences/valid_pad_%d_%d.npy' % (maxlen, num_words))
test_fact_pad_seq = np.load('./variables/pad_sequences/test_pad_%d_%d.npy' % (maxlen, num_words))
train_labels = np.load('./variables/labels/train_one_hot_%s.npy' % (label_type))
valid_labels = np.load('./variables/labels/valid_one_hot_%s.npy' % (label_type))
test_labels = np.load('./variables/labels/test_one_hot_%s.npy' % (label_type))
# label list 标签的类别以及一共有多少类
set_labels = np.load('./variables/label_set/set_%s.npy' % label_type)
##################################################
# model parameter
num_classes = train_labels.shape[1]
num_filters = 512
num_hidden = 1000
batch_size = 256
num_epochs = 2
dropout_rate = 0.2
##################################################
# simple textcnn
input = Input(shape=[train_fact_pad_seq.shape[1]], dtype='float64')
embedding_layer = Embedding(input_dim=num_words + 1,
input_length=maxlen,
output_dim=embedding_dim,
mask_zero=0,
name='Embedding')
embed = embedding_layer(input)
# 词窗大小分别为3,4,5
# filter_size = 3
cnn1 = Conv1D(num_filters, 3, strides=1, padding='same')(embed)
relu1 = Activation(activation='relu')(cnn1)
cnn1 = GlobalMaxPooling1D()(relu1)
# filter_size =4
cnn2 = Conv1D(num_filters, 4, strides=1, padding='same')(embed)
relu2 = Activation(activation='relu')(cnn2)
cnn2 = GlobalMaxPooling1D()(relu2)
# filter_size = 5
cnn3 = Conv1D(num_filters, 5, strides=1, padding='same')(embed)
relu3 = Activation(activation='relu')(cnn3)
cnn3 = GlobalMaxPooling1D()(relu3)
# filter_size = 6
cnn4 = Conv1D(num_filters, 6, strides=1, padding='same')(embed)
relu4 = Activation(activation='relu')(cnn4)
cnn4 = GlobalMaxPooling1D()(relu4)
# 合并三个模型的输出向量
cnn = concatenate([cnn1, cnn2, cnn3, cnn4], axis=-1)
bn = BatchNormalization()(cnn)
drop1 = Dropout(dropout_rate)(bn)
dense = Dense(num_hidden, activation="relu")(drop1)
drop2 = Dropout(dropout_rate)(dense)
main_output = Dense(num_classes, activation='sigmoid')(drop2)
model = Model(inputs=input, outputs=main_output)
model.compile(loss='binary_crossentropy',
optimizer='adam',
metrics=['accuracy'])
plot_model(model, to_file="./pics/textcnn_filter345.png",show_shapes=True)
#########################################################################
for epoch in range(num_epochs):
model.fit(x=train_fact_pad_seq, y=train_labels,
batch_size=batch_size, epochs=1,
validation_data=(valid_fact_pad_seq, valid_labels), verbose=1)
# 计算准确率 calculate accuracy
predictions_valid = model.predict(valid_fact_pad_seq[:]) # 使用valid集做验证
# predictions_test = model.predict(test_fact_pad_seq[:]) # 使用test集做验证
# 用validation集来验证
predictions = predictions_valid
sets = set_labels
y1 = label2tag(valid_labels[:], sets) # 将验证集标签由one-hot转为原标签
y2 = predict2toptag(predictions, sets) # 只取最高的
y3 = predict2half(predictions, sets) # 只取概率大于0.5的
y4 = predict2tag(predictions, sets) # y2与y3的交集，即有大于0.5的取大于0.5的，没有就取概率最高的
# 只取最高置信度的准确率
s1 = [str(y1[i]) == str(y2[i]) for i in range(len(y1))]
print(sum(s1) / len(s1))
# 只取置信度大于0.5的准确率
s2 = [str(y1[i]) == str(y3[i]) for i in range(len(y1))]
print(sum(s2) / len(s2))
# 结合前两个
s3 = [str(y1[i]) == str(y4[i]) for i in range(len(y1))]
accuracy = int(np.round(sum(s3) / len(s3), 3) * 100)
print(accuracy)
# 计算f1 score calculate f1 score
predictions_one_hot = predict1hot(predictions)
f1_micro = f1_score(valid_labels,predictions_one_hot,average='micro')
print('f1_micro_accusation:', f1_micro)
f1_marco = f1_score(valid_labels, predictions_one_hot, average='macro')
print('f1_macro_accusation:', f1_marco)
# 取两者平均
f1_average = int(np.round((f1_marco + f1_micro) / 2, 2) * 100)
print('total:', f1_average)
# # save model
# model.save('./model/textcnn_%s_token_%s_pad_%s_filter_%s_hidden_%s_epoch_%s_accu_%s_f1_%s.h5' % (
# label_type , num_words, maxlen, num_filters, num_hidden, epoch + 1, accuracy, f1_average))
#
# # excel 保存原label与prediction的对比
# r = pd.DataFrame({'label': y1, 'predict': y4})
# r.to_excel('./results/textcnn_%s_token_%s_pad_%s_filter_%s_hidden_%s_epoch_%s_accu_%s_f1_%s.xlsx' % (
# label_type, num_words, maxlen, num_filters, num_hidden, epoch + 1, accuracy, f1_average),
# sheet_name='1', index=False)