-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrnn_autoencoder.py
executable file
·150 lines (124 loc) · 5.23 KB
/
rnn_autoencoder.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
#!/usr/bin/env python3
"""
Training RNN autoencoder
"""
import tensorflow as tf
import numpy as np
import random
import sys
from utils import nn_model
from utils.config import TrainConfig, DecodeConfig
from utils.data_parser import load_data_into_mem, padding
#Data config
scp_file = 'feature_scp/train.scp'
dev_scp_file = 'feature_scp/dev.scp'
max_len = 777
min_epoch = 20
def print_progress(progress, total, loss, output_msg):
sys.stdout.write('\b'*len(output_msg))
new_output_msg = "Progress: {}/{} loss:{:.4f}"\
.format(progress, total, loss)
sys.stdout.write(new_output_msg)
sys.stdout.flush()
return new_output_msg
if __name__ == '__main__':
tr_config = TrainConfig("info")
tr_config.max_len = max_len
tr_config.show_config()
sys.stdout.flush()
max_len = max_len
tr_dropout = tr_config.dropout_keep_prob
tr_zoneout = tr_config.zoneout_keep_prob
max_epoch = tr_config.max_epoch
tr_batch = tr_config.batch_size
print("=============================================================")
print(" Loading data ")
print("=============================================================")
sys.stdout.flush()
data_list = load_data_into_mem(scp_file)
dev_data_list = load_data_into_mem(dev_scp_file)
feature_dim = len(data_list[0][0])
train_utt_num = len(data_list)
dev_utt_num = len(dev_data_list)
print("feature dim: " + str(feature_dim))
print("utt_num: " + str(train_utt_num))
print("dev_utt_num: " + str(dev_utt_num))
tr_config.feature_dim = feature_dim
print("=============================================================")
print(" Set up models ")
print("=============================================================")
sys.stdout.flush()
sess = tf.Session()
#setup nn model's input tensor
x = tf.placeholder(tf.float32, [None, max_len, feature_dim])
y_ = tf.placeholder(tf.float32, [None, max_len, feature_dim])
batch_size = tf.placeholder(tf.int32)
add_noise = tf.placeholder(tf.bool)
dropout_keep_prob = tf.placeholder(tf.float32)
zoneout_keep_prob = tf.placeholder(tf.float32)
##pack the input tensors
input_tensors = {}
input_tensors["x"] = x
input_tensors["y_"] = y_
input_tensors["batch_size"] = batch_size
input_tensors["add_noise"] = add_noise
input_tensors["dropout_keep_prob"] = dropout_keep_prob
input_tensors["zoneout_keep_prob"] = zoneout_keep_prob
model = nn_model.NeuralNetwork(tr_config, input_tensors)
model.setup_train()
model.init_vars(sess)
output_msg = ''
best_dev_recon_loss = float('inf')
print("=============================================================")
print(" Start Training ")
print("=============================================================")
sys.stdout.flush()
for epoch in range(1, max_epoch + 1):
print("[ Epoch {} ]".format(epoch))
print("")
random.shuffle(data_list)
counter = 0
while counter < train_utt_num:
remain_utt_num = train_utt_num - counter
batch_size = min(tr_config.batch_size, remain_utt_num)
X = padding(data_list[counter:counter + batch_size], \
max_len, feature_dim)
counter += len(X)
model.train(sess, X, X, tr_dropout, tr_zoneout, True, len(X))
recon_loss = model.get_tensor_val('re_loss', sess, X, X, len(X))
output_msg = print_progress(counter, train_utt_num, recon_loss,\
output_msg)
print('')
if epoch >= min_epoch:
#dev eval
output_msg = ''
counter = 0
##precision/recall_set dimension: [th * utt]
recon_loss_list = []
while counter < dev_utt_num:
remain_utt_num = dev_utt_num - counter
batch_size = min(tr_config.batch_size, remain_utt_num)
X = padding(dev_data_list[counter:counter + batch_size], \
max_len, feature_dim)
counter += len(X)
recon_loss = model.get_tensor_val('re_loss', sess, X, X, len(X))
recon_loss_list.append(recon_loss)
output_msg = print_progress(counter, dev_utt_num, recon_loss,\
output_msg)
dev_recon_loss = sum(recon_loss_list) / len(recon_loss_list)
print('')
print('The dev loss is {:.4f}'.format(dev_recon_loss))
if epoch == min_epoch:
model.save_vars(sess, tr_config.model_loc)
best_dev_recon_loss = dev_recon_loss
continue
if dev_recon_loss < best_dev_recon_loss:
model.save_vars(sess, tr_config.model_loc)
best_dev_recon_loss = dev_recon_loss
else:
print('Performance get worse, stop training')
break
print('=============================================================')
print(' Training finished, show config info. again ')
print('=============================================================')
tr_config.show_config()