/
models.py
213 lines (164 loc) · 7.59 KB
/
models.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
import tensorflow as tf
from ops import lrelu, batch_norm, linear, conv2d, maxpool2d, fc, layer_norm, upsample2x, downsample2x
"""
2 type models are defined:
- DCGAN [Radford+, ICLR2016]
- ResNet [He+, CVPR2016]
- This model needs a lot of GPU memory, so you should reduce batch size.
"""
class GeneratorDCGAN():
def __init__(self, img_size=(128, 128), img_dim=1, z_size=100, k_size=5, layer_n=3,
smallest_hidden_unit_n=128, is_bn=True):
self.img_size = img_size
self.img_dim = img_dim
self.z_size = z_size
self.k_size = k_size
self.layer_n = layer_n
self.smallest_hidden_unit_n = smallest_hidden_unit_n
self.is_bn = is_bn
def __call__(self, x, is_reuse=False, is_train=True):
with tf.variable_scope('generator') as scope:
if is_reuse:
scope.reuse_variables()
unit_size = self.img_size[0] // (2 ** self.layer_n)
unit_n = self.smallest_hidden_unit_n * (2 ** (self.layer_n - 1))
batch_size = int(x.shape[0])
with tf.variable_scope('pre'):
x = linear(x, unit_size * unit_size * unit_n)
x = tf.reshape(x, (batch_size, unit_size, unit_size, unit_n))
if self.is_bn:
x = batch_norm(x, is_train)
x = tf.nn.relu(x)
for i in range(self.layer_n):
with tf.variable_scope('layer{}'.format(i)):
if i == self.layer_n - 1:
unit_n = self.img_dim
else:
unit_n = self.smallest_hidden_unit_n * (2 ** (self.layer_n - i - 2))
x_shape = x.get_shape().as_list()
x = tf.image.resize_bilinear(x, (x_shape[1] * 2, x_shape[2] * 2))
x = conv2d(x, unit_n, self.k_size, 1, 'SAME')
if i != self.layer_n - 1:
if self.is_bn:
x = batch_norm(x, is_train)
x = tf.nn.relu(x)
x = tf.nn.tanh(x)
return x
class DiscriminatorDCGAN():
def __init__(self, img_size=(128, 128), img_dim=1, k_size=5, layer_n=3, smallest_hidden_unit_n=128, is_bn=True):
self.img_size = img_size
self.img_dim = img_dim
self.k_size = k_size
self.layer_n = layer_n
self.smallest_hidden_unit_n = smallest_hidden_unit_n
self.is_bn = is_bn
def __call__(self, x, is_reuse=False, is_train=True):
with tf.variable_scope('discriminator') as scope:
if is_reuse:
scope.reuse_variables()
unit_n = self.smallest_hidden_unit_n
batch_size = int(x.shape[0])
for i in range(self.layer_n):
with tf.variable_scope('layer{}'.format(i + 1)):
x = conv2d(x, unit_n, self.k_size, 2, 'SAME')
if self.is_bn and i != 0:
x = batch_norm(x, is_train)
x = lrelu(x)
unit_n = self.smallest_hidden_unit_n * (2 ** (i + 1))
x = tf.reshape(x, (batch_size, -1))
x = linear(x, 1)
return x
class GeneratorResNet():
def __init__(self, k_size=3, smallest_unit_n=64):
self.k_size = k_size
self.smallest_unit_n = smallest_unit_n
def _residual_block(self, x, n_out, is_train, name='residual'):
with tf.variable_scope(name):
with tf.variable_scope('shortcut'):
x1 = upsample2x(x)
x1 = conv2d(x1, n_out, self.k_size, 1, 'SAME')
with tf.variable_scope('normal'):
x2 = batch_norm(x, is_train, name='batch_norm_0')
x2 = tf.nn.relu(x2)
x2 = upsample2x(x2)
x2 = conv2d(x2, n_out, self.k_size, 1, 'SAME', name='conv2d_0')
x2 = batch_norm(x2, is_train, name='batch_norm_1')
x2 = tf.nn.relu(x2)
x2 = conv2d(x2, n_out, self.k_size, 1, 'SAME', name='conv2d_1')
return x1 + x2
def __call__(self, x, is_train=True, is_reuse=False):
with tf.variable_scope('generator') as scope:
if is_reuse:
scope.reuse_variables()
with tf.variable_scope('first'):
x = linear(x, 4 * 4 * 8 * self.smallest_unit_n)
x = tf.reshape(x, [-1, 4, 4, 8 * self.smallest_unit_n])
for i, times in enumerate([8, 4, 2, 1]):
x = self._residual_block(x, times * self.smallest_unit_n, is_train, 'residual_{}'.format(i))
with tf.variable_scope('last'):
x = batch_norm(x, is_train)
x = tf.nn.relu(x)
x = conv2d(x, 3, self.k_size, 1, 'SAME')
x = tf.tanh(x)
return x
class DiscriminatorResNet():
def __init__(self, k_size=3, smallest_unit_n=64):
self.k_size = k_size
self.smallest_unit_n = smallest_unit_n
def _residual_block(self, x, n_out, name='residual'):
with tf.variable_scope(name):
with tf.variable_scope('shortcut'):
x1 = downsample2x(x)
x1 = conv2d(x1, n_out, self.k_size, 1, 'SAME')
with tf.variable_scope('normal'):
x2 = layer_norm(x, name='layer_norm_0')
x2 = tf.nn.relu(x2)
x2 = conv2d(x2, n_out, self.k_size, 1, 'SAME', name='conv2d_0')
x2 = layer_norm(x2, name='layer_norm_1')
x2 = tf.nn.relu(x2)
x2 = downsample2x(x2)
x2 = conv2d(x2, n_out, self.k_size, 1, 'SAME', name='conv2d_1')
return x1 + x2
def __call__(self, x, is_train=True, is_reuse=False):
with tf.variable_scope('discriminator') as scope:
if is_reuse:
scope.reuse_variables()
with tf.variable_scope('first'):
x = conv2d(x, self.smallest_unit_n, self.k_size, 1, 'SAME')
for i, times in enumerate([2, 4, 8, 8]):
x = self._residual_block(x, times * self.smallest_unit_n, 'residual_{}'.format(i))
x = tf.reshape(x, [-1, 4 * 4 * 8 * self.smallest_unit_n])
with tf.variable_scope('last'):
x = linear(x, 1)
return x
class Classifier():
def __init__(self, img_size, img_dim, k_size, class_n, smallest_unit_n=64):
self.img_size = img_size
self.img_dim = img_dim
self.k_size = k_size
self.class_n = class_n
self.smallest_unit_n = smallest_unit_n
def __call__(self, x, is_reuse=False, is_train=True):
with tf.variable_scope('classifier') as scope:
if is_reuse:
scope.reuse_variables()
unit_n = self.smallest_unit_n
conv_ns = [2, 2]
for layer_i, conv_n in enumerate(conv_ns):
with tf.variable_scope('layer{}'.format(layer_i)):
for conv_i in range(conv_n):
x = conv2d(x, unit_n, self.k_size, 1, 'SAME', name='conv2d_{}'.format(conv_i))
x = tf.nn.relu(x)
x = maxpool2d(x, self.k_size, 2, 'SAME')
unit_n *= 2
unit_n = 256
fc_n = 1
for layer_i in range(len(conv_ns), len(conv_ns) + fc_n):
with tf.variable_scope('layer{}'.format(layer_i)):
x = fc(x, unit_n)
x = tf.nn.relu(x)
x = batch_norm(x, is_train)
x = tf.nn.dropout(x, 0.5)
with tf.variable_scope('output'.format(layer_i)):
x = fc(x, self.class_n)
return x