forked from CSTR-Edinburgh/ophelia
-
Notifications
You must be signed in to change notification settings - Fork 8
/
modules.py
263 lines (221 loc) · 9.4 KB
/
modules.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
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
# -*- coding: utf-8 -*-
#/usr/bin/python2
'''
By kyubyong park. [email protected].
https://www.github.com/kyubyong/dc_tts
Modified...
'''
from __future__ import print_function, division
import tensorflow as tf
from logging import info
def embed(inputs, vocab_size, num_units, zero_pad=True, scope="embedding", reuse=None):
'''Embeds a given tensor.
Args:
inputs: A `Tensor` with type `int32` or `int64` containing the ids
to be looked up in `lookup table`.
vocab_size: An int. Vocabulary size.
num_units: An int. Number of embedding hidden units.
zero_pad: A boolean. If True, all the values of the fist row (id 0)
should be constant zeros.
scope: Optional scope for `variable_scope`.
reuse: Boolean, whether to reuse the weights of a previous layer
by the same name.
Returns:
A `Tensor` with one more rank than inputs's. The last dimensionality
should be `num_units`.
'''
with tf.variable_scope(scope, reuse=reuse):
lookup_table = tf.get_variable('lookup_table',
dtype=tf.float32,
shape=[vocab_size, num_units],
initializer=tf.truncated_normal_initializer(mean=0.0, stddev=0.1))
if zero_pad:
lookup_table = tf.concat((tf.zeros(shape=[1, num_units]),
lookup_table[1:, :]), 0)
outputs = tf.nn.embedding_lookup(lookup_table, inputs)
return outputs
def normalize(inputs,
scope="normalize",
reuse=None,
normtype='layer'):
'''Applies layer normalization that normalizes along the last axis.
Args:
inputs: A tensor with 2 or more dimensions, where the first dimension has
`batch_size`. The normalization is over the last dimension.
scope: Optional scope for `variable_scope`.
reuse: Boolean, whether to reuse the weights of a previous layer
by the same name.
Returns:
A tensor with the same shape and data dtype as `inputs`.
'''
assert normtype in [None, 'layer', 'batch']
if normtype == 'layer':
outputs = tf.contrib.layers.layer_norm(inputs,
begin_norm_axis=-1,
scope=scope,
reuse=reuse)
elif normtype == 'batch':
outputs = tf.contrib.layers.batch_norm(inputs,
scope=scope,
reuse=reuse)
else:
outputs = inputs
return outputs
def learn_channel_contributions(input_tensor, codes, ncodes=1, reuse=None):
# codes (B, ?) Always 1D??
info('learn_channel_contributions; codes: %s'%(codes.shape))
nchannels = input_tensor.get_shape().as_list()[-1]
lcc_gate = embed(codes, vocab_size=ncodes, num_units=nchannels, \
scope="lcc_embed", reuse=reuse) ## init weight mean 0.0
lcc_gate = tf.nn.sigmoid(lcc_gate, "lcc_gate") ## -> 0.5 after sigmoid
## lcc_gate (B, filters)
#print(lcc_gate.shape) # (32, ?, 512)
input_tensor = lcc_gate * input_tensor ## Broadcast on time dimension
return input_tensor
def conv1d(inputs,
filters=None,
size=1,
rate=1,
padding="SAME",
dropout_rate=0,
use_bias=True,
activation_fn=None,
training=True,
scope="conv1d",
reuse=None,
normtype='layer',
lcc=0, codes=None):
'''
Args:
inputs: A 3-D tensor with shape of [batch, time, depth].
filters: An int. Number of outputs (=activation maps)
size: An int. Filter size.
rate: An int. Dilation rate.
padding: Either `same` or `valid` or `causal` (case-insensitive).
dropout_rate: A float of [0, 1].
use_bias: A boolean.
activation_fn: A string.
training: A boolean. If True, dropout is applied.
scope: Optional scope for `variable_scope`.
reuse: Boolean, whether to reuse the weights of a previous layer
by the same name.
Returns:
A masked tensor of the same shape and dtypes as `inputs`.
'''
with tf.variable_scope(scope):
if padding.lower() == "causal":
# pre-padding for causality
pad_len = (size - 1) * rate # padding size
inputs = tf.pad(inputs, [[0, 0], [pad_len, 0], [0, 0]])
padding = "valid"
if filters is None:
filters = inputs.get_shape().as_list()[-1]
params = {"inputs": inputs, "filters": filters, "kernel_size": size,
"dilation_rate": rate, "padding": padding, "use_bias": use_bias,
"kernel_initializer": tf.contrib.layers.variance_scaling_initializer(), "reuse": reuse}
tensor = tf.layers.conv1d(**params)
tensor = normalize(tensor, normtype=normtype, reuse=reuse)
if activation_fn is not None:
tensor = activation_fn(tensor)
tensor = tf.layers.dropout(tensor, rate=dropout_rate, training=training)
if lcc:
tensor = learn_channel_contributions(tensor, codes, ncodes=lcc, reuse=reuse)
return tensor
def hc(inputs,
filters=None,
size=1,
rate=1,
padding="SAME",
dropout_rate=0,
use_bias=True,
activation_fn=None,
training=True,
scope="hc",
reuse=None,
normtype='layer', lcc=0, codes=None):
'''
Args:
inputs: A 3-D tensor with shape of [batch, time, depth].
filters: An int. Number of outputs (=activation maps)
size: An int. Filter size.
rate: An int. Dilation rate.
padding: Either `same` or `valid` or `causal` (case-insensitive).
use_bias: A boolean.
activation_fn: A string.
training: A boolean. If True, dropout is applied.
scope: Optional scope for `variable_scope`.
reuse: Boolean, whether to reuse the weights of a previous layer
by the same name.
Returns:
A masked tensor of the same shape and dtypes as `inputs`.
'''
_inputs = inputs
with tf.variable_scope(scope):
if padding.lower() == "causal":
# pre-padding for causality
pad_len = (size - 1) * rate # padding size
inputs = tf.pad(inputs, [[0, 0], [pad_len, 0], [0, 0]])
padding = "valid"
if filters is None:
filters = inputs.get_shape().as_list()[-1]
params = {"inputs": inputs, "filters": 2*filters, "kernel_size": size,
"dilation_rate": rate, "padding": padding, "use_bias": use_bias,
"kernel_initializer": tf.contrib.layers.variance_scaling_initializer(), "reuse": reuse}
tensor = tf.layers.conv1d(**params)
H1, H2 = tf.split(tensor, 2, axis=-1)
H1 = normalize(H1, scope="H1", normtype=normtype, reuse=reuse)
H2 = normalize(H2, scope="H2", normtype=normtype, reuse=reuse)
H1 = tf.nn.sigmoid(H1, "gate")
H2 = activation_fn(H2, "info") if activation_fn is not None else H2
if lcc: ## LCC applied on transformation connections only
H2 = learn_channel_contributions(H2, codes, ncodes=lcc, reuse=reuse)
tensor = H1*H2 + (1.-H1)*_inputs
tensor = tf.layers.dropout(tensor, rate=dropout_rate, training=training)
return tensor
def conv1d_transpose(inputs,
filters=None,
size=3,
stride=2,
padding='same',
dropout_rate=0,
use_bias=True,
activation=None,
training=True,
scope="conv1d_transpose",
reuse=None,
normtype='layer'):
'''
Args:
inputs: A 3-D tensor with shape of [batch, time, depth].
filters: An int. Number of outputs (=activation maps)
size: An int. Filter size.
rate: An int. Dilation rate.
padding: Either `same` or `valid` or `causal` (case-insensitive).
dropout_rate: A float of [0, 1].
use_bias: A boolean.
activation_fn: A string.
training: A boolean. If True, dropout is applied.
scope: Optional scope for `variable_scope`.
reuse: Boolean, whether to reuse the weights of a previous layer
by the same name.
Returns:
A tensor of the shape with [batch, time*2, depth].
'''
with tf.variable_scope(scope, reuse=reuse):
if filters is None:
filters = inputs.get_shape().as_list()[-1]
inputs = tf.expand_dims(inputs, 1)
tensor = tf.layers.conv2d_transpose(inputs,
filters=filters,
kernel_size=(1, size),
strides=(1, stride),
padding=padding,
activation=None,
kernel_initializer=tf.contrib.layers.variance_scaling_initializer(),
use_bias=use_bias)
tensor = tf.squeeze(tensor, 1)
tensor = normalize(tensor, normtype=normtype, reuse=reuse)
if activation is not None:
tensor = activation(tensor)
tensor = tf.layers.dropout(tensor, rate=dropout_rate, training=training)
return tensor