-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
180 lines (159 loc) · 6.23 KB
/
main.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
# -*- coding: utf-8 -*-
# requirement:python3
# Author: binglel
# Last modified: 20181114 11:00
# Email: chyb3.14@gmail.com
# Filename: main
# Description:
# ******************************************************
import time
import numpy as np
import tensorflow as tf
import scipy.io.wavfile as wav
from python_speech_features import mfcc
from utils import maybe_download as maybe_download
from utils import sparse_tuple_from as sparse_tuple_from
# Constants
SPACE_TOKEN = '<space>'
SPACE_INDEX = 0
# 0 is reserved to space
FIRST_INDEX = ord('a') - 1
print (FIRST_INDEX)
# Some configs
num_features = 13
# Number of units in the LSTM cell
num_units=50
num_classes =26 + 1 + 1
num_epochs = 120
num_hidden = 50
num_layers = 1
batch_size = 1
initial_learning_rate = 1e-2
momentum = 0.9
num_examples = 1
num_batches_per_epoch = int(num_examples/batch_size)
audio_filename = maybe_download('LDC93S1.wav', 93638)
target_filename = maybe_download('LDC93S1.txt', 62)
# fs is framerate
fs, audio = wav.read(audio_filename)
inputs = mfcc(audio, samplerate=fs)
print (inputs.shape)
# Tranform in 3D array
train_inputs = np.asarray(inputs[np.newaxis, :])
print (train_inputs.shape)
# Normalize
train_inputs = (train_inputs - np.mean(train_inputs))/np.std(train_inputs)
print (train_inputs.shape)
train_seq_len = [train_inputs.shape[1]]
with open(target_filename, 'r') as f:
#Only the last line is necessary
line = f.readlines()[-1]
# Get only the words between [a-z] and replace period for none
original = ' '.join(line.strip().lower().split(' ')[2:]).replace('.', '')
targets = original.replace(' ', ' ')
targets = targets.split(' ')
print (targets)
targets = np.hstack([SPACE_TOKEN if x == '' else list(x) for x in targets])
print (targets)
# Transform char into index
targets = np.asarray([SPACE_INDEX if x == SPACE_TOKEN else ord(x) - FIRST_INDEX for x in targets])
print (targets)
# Creating sparse representation to feed the placeholder
train_targets = sparse_tuple_from([targets])
print (train_targets)
# We don't have a validation dataset :(
val_inputs, val_targets, val_seq_len = train_inputs, train_targets, train_seq_len
graph = tf.Graph()
with graph.as_default():
'''
inputs是输入的placeholder
输入的尺寸是[batch_size, max_stepsize, num_features],
but the batch_size and max_stepsize can vary along each step
'''
inputs = tf.placeholder(tf.float32, [None, None, num_features])
# Here we use sparse_placeholder that will generate a SparseTensor required by ctc_loss op.
targets = tf.sparse_placeholder(tf.int32)
# 1d array of size [batch_size]
seq_len = tf.placeholder(tf.int32, [None])
with graph.as_default():
'''
Defining the cell
Can be:
tf.nn.rnn_cell.RNNCell
tf.nn.rnn_cell.GRUCell
'''
cells = []
for _ in range(num_layers):
cell = tf.contrib.rnn.LSTMCell(num_units) # Or LSTMCell(num_units)
cells.append(cell)
stack = tf.contrib.rnn.MultiRNNCell(cells)
# The second output is the last state and we will no use that
outputs, _ = tf.nn.dynamic_rnn(stack, inputs, seq_len, dtype=tf.float32)
print (outputs)
shape = tf.shape(inputs)
print (shape)
batch_s, max_timesteps = shape[0], shape[1]
print (batch_s)
print (max_timesteps)
# Reshaping to apply the same weights over the timesteps
outputs = tf.reshape(outputs, [-1, num_hidden])
print (outputs)
# Truncated normal with mean 0 and stdev=0.1
# Tip: Try another initialization
# see https://www.tensorflow.org/versions/r0.9/api_docs/python/contrib.layers.html#initializers
W = tf.Variable(tf.truncated_normal([num_hidden,
num_classes],
stddev=0.1))
# Zero initialization
# Tip: Is tf.zeros_initializer the same?
b = tf.Variable(tf.constant(0., shape=[num_classes]))
# Doing the affine projection
logits = tf.matmul(outputs, W) + b
# Reshaping back to the original shape
logits = tf.reshape(logits, [batch_s, -1, num_classes])
# Time major
logits = tf.transpose(logits, (1, 0, 2))
# Option 2:
# (it's slower but you'll get better results)
print(logits.get_shape())
with graph.as_default():
loss = tf.nn.ctc_loss(targets, logits, seq_len)
cost = tf.reduce_mean(loss)
optimizer = tf.train.MomentumOptimizer(initial_learning_rate, 0.9).minimize(cost)
# decoder style
with graph.as_default():
# decoded, log_prob = tf.nn.ctc_greedy_decoder(logits, seq_len)
decoded, log_prob = tf.nn.ctc_beam_search_decoder(logits, seq_len)
with graph.as_default():
ler = tf.reduce_mean(tf.edit_distance(tf.cast(decoded[0], tf.int32), targets))
session=tf.Session(graph=graph)
with graph.as_default():
init=tf.global_variables_initializer()
session.run(init)
for curr_epoch in range(num_epochs):
train_cost = train_ler = 0
start = time.time()
for batch in range(num_batches_per_epoch):
feed = {inputs: train_inputs, targets: train_targets, seq_len: train_seq_len}
batch_cost, _ = session.run([cost, optimizer], feed)
train_cost += batch_cost*batch_size
train_ler += session.run(ler, feed_dict=feed)*batch_size
train_cost /= num_examples
train_ler /= num_examples
val_feed = {inputs: val_inputs, targets: val_targets, seq_len: val_seq_len}
val_cost, val_ler = session.run([cost, ler], feed_dict=val_feed)
log = "Epoch {}/{}, train_cost = {:.3f}, train_ler = {:.3f}, val_cost = {:.3f}, val_ler = {:.3f}, time = {:.3f}"
print(log.format(curr_epoch+1, num_epochs, train_cost, train_ler, val_cost, val_ler, time.time() - start))
feed = {inputs: train_inputs,targets: train_targets,seq_len: train_seq_len}
d = session.run(decoded[0], feed_dict=val_feed)
print(d)
str_decoded = ''.join([chr(x) for x in np.asarray(d[1]) + FIRST_INDEX])
print(str_decoded)
# Replacing blank label to none
str_decoded = str_decoded.replace(chr(ord('z') + 1), '')
print(str_decoded)
# Replacing space label to space
str_decoded = str_decoded.replace(chr(ord('a') - 1), ' ')
print('Original:\n%s' % original)
print('Decoded:\n%s' % str_decoded)
session.close()