-
Notifications
You must be signed in to change notification settings - Fork 3
/
name_model.py
148 lines (116 loc) · 4.67 KB
/
name_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
from utils import *
import tensorflow as tf
import functools
from keras import backend as K
from keras.models import load_model
from keras.models import Sequential
from keras.layers import Dense, LSTM, Bidirectional
from keras.callbacks import ModelCheckpoint
from keras.optimizers import Adam
import matplotlib.pyplot as plt
def as_keras_metric(method):
@functools.wraps(method)
def wrapper(self, args, **kwargs):
""" Wrapper for turning tensorflow metrics into keras metrics """
value, update_op = method(self, args, **kwargs)
K.get_session().run(tf.local_variables_initializer())
with tf.control_dependencies([update_op]):
value = tf.identity(value)
return value
return wrapper
class keras_Model:
def __init__(self):
self.model = None
self.model_path = './model/name_model.h5'
def load(self):
self.model = load_model(
'./model/name_model.h5', custom_objects={
'auc': as_keras_metric(tf.metrics.auc)
}
)
def pred(self, seq):
return self.model.predict(
name_one_hot(seq, 15).reshape(1, 15, 26)
)[0]
def build_model(self):
model = Sequential()
model.add(Bidirectional(LSTM(64, input_shape=(15, 26))))
model.add(Dense(3, activation='softmax', kernel_initializer='normal'))
model.compile(loss='categorical_crossentropy', optimizer=Adam(0.0005),
metrics=['accuracy', as_keras_metric(tf.metrics.auc)])
self.model = model
def show_train_graph(self, hist):
fig, loss_ax = plt.subplots()
acc_ax = loss_ax.twinx()
loss_ax.plot(hist.history['loss'], 'y', label='train loss')
loss_ax.plot(hist.history['val_loss'], 'r', label='val loss')
acc_ax.plot(hist.history['acc'], 'b', label='train acc')
acc_ax.plot(hist.history['val_acc'], 'g', label='val acc')
acc_ax.plot(hist.history['auc'], 'm', label='train auc')
acc_ax.plot(hist.history['val_auc'], 'k', label='val auc')
loss_ax.set_xlabel('epoch')
loss_ax.set_ylabel('loss')
acc_ax.set_ylabel('auc_roc')
loss_ax.legend(loc='upper left')
acc_ax.legend(loc='lower left')
plt.show()
fig.savefig('./model/train_graph.png')
def train(self):
# ------------------------------------
max_seq_len = 15
np.random.seed(5)
# ------------------------------------
kr_list = get_file('./data/kr_first_names.txt')
ch_list = get_file('./data/ch_first_names.txt')
us_list = get_file('./data/us_first_names.txt')
a = len(kr_list)
b = len(ch_list)
c = len(us_list)
data_len = a + b + c
X, Y = [], []
for _ in range(1):
for name in kr_list:
X.append(name_one_hot(name, max_seq_len))
Y.append(np.array([1, 0, 0]))
for _ in range(1):
for name in ch_list:
X.append(name_one_hot(name, max_seq_len))
Y.append(np.array([0, 1, 0]))
for name in us_list:
X.append(name_one_hot(name, max_seq_len))
Y.append(np.array([0, 0, 1]))
X, Y = np.array(X), np.array(Y)
np.reshape(X, (data_len, max_seq_len, 26))
np.reshape(Y, (data_len, 1, 3))
permutation = np.random.permutation(X.shape[0])
X = X[permutation]
Y = Y[permutation]
train_len = int(data_len * 0.99)
x_train = X[:train_len]
y_train = Y[:train_len]
x_val = X[train_len:]
y_val = Y[train_len:]
loss_CP = ModelCheckpoint(
'./model/loss.h5', monitor='val_loss', mode='min',
verbose=0, save_best_only=True
)
acc_CP = ModelCheckpoint(
'./model/acc.h5', monitor='val_acc', mode='max',
verbose=0, save_best_only=True
)
auc_CP = ModelCheckpoint(
'./model/auc.h5', monitor='val_auc', mode='max',
verbose=0, save_best_only=True
)
self.build_model()
model = self.model
hist = model.fit(x_train, y_train, epochs=300, batch_size=512,
validation_data=(x_val, y_val), verbose=2,
callbacks=[loss_CP, acc_CP, auc_CP])
# score = model.evaluate(x_test, y_test)
# print("%s: %.2f%%" %(model.metrics_names[1], score[1] * 100))
# model.save(self.model_path)
self.show_train_graph(hist)
if __name__=='__main__':
km = keras_Model()
km.train()