def discriminator(self, x, y, do_rate=0.5, reuse=None):
with tf.variable_scope("discriminator", reuse=reuse):
"""
x = t.dense(x, self.fc_unit * 5, name='disc-fc-x')
x = tf.reshape(x, (-1, self.fc_unit, 5))
x = tf.reduce_max(x, axis=-1, keepdims=False, name='disc-maxout-x')
x = tf.nn.relu(x)
x = tf.layers.dropout(x, do_rate, name='disc-do-x')
y = t.dense(y, (self.fc_unit // 4) * 5, name='disc-fc-y')
y = tf.reshape(y, (-1, (self.fc_unit // 4), 5))
y = tf.reduce_max(y, axis=-1, keepdims=False, name='disc-maxout-y')
y = tf.nn.relu(y)
y = tf.layers.dropout(y, do_rate, name='disc-do-y')
"""
x = tf.concat([x, y], axis=1)
x = t.dense(x, self.fc_unit * 5, name='disc-fc-1')
x = tf.reshape(x, (-1, self.fc_unit, 5))
x = tf.reduce_max(x, axis=-1, keepdims=False, name='disc-maxout-1')
x = tf.nn.relu(x)
x = tf.layers.dropout(x, do_rate, name='disc-do-1')
x = t.dense(x, self.fc_unit * 4, name='disc-fc-2')
x = tf.reshape(x, (-1, self.fc_unit, 4))
x = tf.reduce_max(x, axis=-1, keepdims=False, name='disc-maxout-2')
x = tf.nn.relu(x)
x = tf.layers.dropout(x, do_rate, name='disc-do-2')
x = t.dense(x, 1, name='disc-fc-3')
x = tf.sigmoid(x)
return x
def discriminator(self, x, reuse=None):
"""
# Following a network architecture referred in the paper
:param x: Input images (-1, 384, 384, 3)
:param reuse: re-usability
:return: HR (High Resolution) or SR (Super Resolution) images
"""
with tf.variable_scope("discriminator", reuse=reuse):
x = t.conv2d(x, self.df_dim, 3, 1, name='n64s1-1')
x = tf.nn.leaky_relu(x)
strides = [2, 1]
filters = [1, 2, 2, 4, 4, 8, 8]
for i, f in enumerate(filters):
x = t.conv2d(x,
f=f,
k=3,
s=strides[i % 2],
name='n%ds%d-%d' % (f, strides[i % 2], i + 1))
x = t.batch_norm(x, name='n%d-bn-%d' % (f, i + 1))
x = tf.nn.leaky_relu(x)
x = tf.layers.flatten(x) # (-1, 96 * 96 * 64)
x = t.dense(x, 1024, name='disc-fc-1')
x = tf.nn.leaky_relu(x)
x = t.dense(x, 1, name='disc-fc-2')
# x = tf.nn.sigmoid(x)
return x
def discriminator(self, x, y=None, share_params=False, reuse=False, name=""):
with tf.variable_scope("discriminator-%s" % name, reuse=reuse):
if y is None:
x = tf.layers.flatten(x)
x = tf.concat([x, y], axis=1)
x = t.dense(x, self.height * self.width * self.channel,
name='disc-' + name + '-dense-0-y')
x = tf.reshape(x, self.image_shape)
else:
pass
# Using conv2d pooling instead of max_pool2d because of the speed.
# In the CoGAN paper, max_pool2d is used.
x = t.conv2d(x, f=self.df_dim, k=5, s=2, reuse=False, name='disc-' + name + '-conv2d-0')
x = t.prelu(x, reuse=False, name='disc-' + name + '-prelu-0')
# x = tf.nn.max_pool(x, ksize=2, strides=2, padding='SAME', name='disc' + name + '-max_pool2d-0')
x = t.conv2d(x, f=self.df_dim * 2, k=5, s=2, reuse=False, name='disc-' + name + '-conv2d-1')
x = t.batch_norm(x, is_train=False, name='disc-bn-0')
x = t.prelu(x, reuse=False, name='disc-' + name + '-prelu-1')
# x = tf.nn.max_pool(x, ksize=2, strides=2, padding='SAME', name='disc' + name + '-max_pool2d-1')
x = tf.layers.flatten(x)
x = t.dense(x, self.fc_unit, reuse=share_params, name='disc-dense-0')
x = t.batch_norm(x, is_train=share_params, name='disc-bn-1')
x = t.prelu(x, reuse=share_params, name='disc-prelu-2')
x = t.dense(x, 1, reuse=share_params, name='disc-dense-1')
return x
def discriminator(self, x, reuse=None):
"""
:param x: images
:param reuse: re-usable
:return: logits
"""
with tf.variable_scope("discriminator", reuse=reuse):
x = t.conv2d(x, self.df_dim * 1, 4, 2, name='disc-conv2d-1')
x = tf.nn.leaky_relu(x, alpha=0.1)
x = t.conv2d(x, self.df_dim * 2, 4, 2, name='disc-conv2d-2')
x = t.batch_norm(x, name='disc-bn-1')
x = tf.nn.leaky_relu(x, alpha=0.1)
x = t.conv2d(x, self.df_dim * 4, 4, 2, name='disc-conv2d-3')
x = t.batch_norm(x, name='disc-bn-2')
x = tf.nn.leaky_relu(x, alpha=0.1)
x = t.conv2d(x, self.df_dim * 8, 4, 2, name='disc-conv2d-4')
x = t.batch_norm(x, name='disc-bn-3')
x = tf.nn.leaky_relu(x, alpha=0.1)
x = tf.layers.flatten(x)
x = t.dense(x, self.fc_unit, name='disc-fc-1')
x = t.batch_norm(x, name='disc-bn-4')
x = tf.nn.leaky_relu(x, alpha=0.1)
x = t.dense(x, 1 + self.n_cont + self.n_cat, name='disc-fc-2')
prob, cont, cat = x[:, 0], x[:, 1:1 + self.n_cont], x[:, 1 + self.n_cont:] # logits
prob = tf.nn.sigmoid(prob) # probability
cat = tf.nn.softmax(cat) # categories
return prob, cont, cat
def discriminator(self, x, reuse=None):
"""
# Following a D Network, CiFar-like-hood, referred in the paper
:param x: images
:param y: labels
:param reuse: re-usable
:return: classification, probability (fake or real), network
"""
with tf.variable_scope("discriminator", reuse=reuse):
x = t.conv2d(x, self.df_dim, 3, 2, name='disc-conv2d-1')
x = tf.nn.leaky_relu(x, alpha=0.2)
x = tf.layers.dropout(x, 0.5, name='disc-dropout2d-1')
for i in range(5):
x = t.conv2d(x,
self.df_dim * (2**(i + 1)),
k=3,
s=(i % 2 + 1),
name='disc-conv2d-%d' % (i + 2))
x = t.batch_norm(x, reuse=reuse, name="disc-bn-%d" % (i + 1))
x = tf.nn.leaky_relu(x, alpha=0.2)
x = tf.layers.dropout(x,
0.5,
name='disc-dropout2d-%d' % (i + 1))
net = tf.layers.flatten(x)
cat = t.dense(net, self.n_classes, name='disc-fc-cat')
disc = t.dense(net, 1, name='disc-fc-disc')
return cat, disc, net
def discriminator(self, x, reuse=None):
with tf.variable_scope("discriminator", reuse=reuse):
x = t.dense(x, self.fc_unit, name='disc-fc-1')
x = tf.nn.leaky_relu(x)
x = t.dense(x, 1, name='discd-fc-2')
return x
def generator(self, z, reuse=None):
with tf.variable_scope("generator", reuse=reuse):
x = t.dense(z, self.fc_unit, name='gen-fc-1')
x = tf.nn.leaky_relu(x)
x = t.dense(x, self.n_input, name='gen-fc-2')
x = tf.nn.sigmoid(x)
return x
def discriminator(self, x, reuse=None):
with tf.variable_scope("discriminator", reuse=reuse):
for i in range(2):
x = t.dense(x, self.fc_unit, name='d-fc-%d' % i)
x = tf.nn.leaky_relu(x)
logits = t.dense(x, 1, name='d-fc-2')
prob = tf.nn.sigmoid(logits)
return prob, logits
def generator(self, x, reuse=None, is_train=True):
with tf.variable_scope("generator", reuse=reuse):
for i in range(2):
x = t.dense(x, self.fc_unit, name='g-fc-%d' % i)
x = t.batch_norm(x, is_train=is_train)
x = tf.nn.leaky_relu(x)
logits = t.dense(x, self.n_input, name='g-fc-2')
prob = tf.nn.sigmoid(logits)
return prob
def discriminator(self, x, reuse=None):
with tf.variable_scope('discriminator', reuse=reuse):
x = t.dense(x, self.dfc_unit, name='disc-fc-1')
x = tf.nn.elu(x)
x = t.dense(x, self.dfc_unit, name='disc-fc-2')
x = tf.nn.elu(x)
x = tf.layers.flatten(x)
x = t.dense(x, 1, name='disc-fc-3')
return x
def generator(self, z, y, reuse=None, is_train=True):
with tf.variable_scope("generator", reuse=reuse):
x = tf.concat([z, y], axis=1)
x = t.dense(x, self.fc_unit, name='g-fc-1')
x = tf.nn.dropout(x, .5) if is_train else tf.nn.dropout(x, 1.)
x = tf.nn.leaky_relu(x)
x = t.dense(x, self.n_input, name='g-fc-2')
x = tf.nn.sigmoid(x)
return x
def classifier(self, x, reuse=None):
"""
# Following a C Network, CiFar-like-hood, referred in the paper
:param x: image, shape=(-1, 28, 28, 1)
:param reuse: re-usable
:return: logits
"""
with tf.variable_scope("classifier", reuse=reuse):
x = t.dense(x, self.fc_unit, name='c-fc-1')
x = tf.nn.leaky_relu(x)
x = t.dense(x, self.n_classes, name='c-fc-2')
return x
def generator(self, z, reuse=None, is_train=True):
with tf.variable_scope('generator', reuse=reuse):
x = t.dense(z, self.gfc_unit, name='gen-fc-1')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-1')
x = tf.nn.relu(x)
x = t.dense(x, self.gfc_unit, name='gen-fc-2')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-2')
x = tf.nn.relu(x)
x = t.dense(x, self.n_input, name='gen-fc-3')
x = tf.nn.sigmoid(x)
return x
def generator(self, z, y, do_rate=0.5, reuse=None):
with tf.variable_scope("generator", reuse=reuse):
x = tf.concat([z, y], axis=1)
x = t.dense(x, self.fc_unit * 1, name='gen-fc-1')
x = tf.nn.relu(x)
x = tf.layers.dropout(x, do_rate, name='gen-do-1')
x = t.dense(x, self.fc_unit * 4, name='gen-fc-2')
x = tf.nn.relu(x)
x = tf.layers.dropout(x, do_rate, name='gen-do-2')
x = t.dense(x, self.n_input, name='gen-fc-3')
x = tf.sigmoid(x)
return x
def discriminator(self, x, y, reuse=None):
with tf.variable_scope("discriminator", reuse=reuse):
x = tf.concat([x, y], axis=1)
x = t.dense(x, self.maxout_unit * self.fc_unit, name='d-fc-1')
x = tf.reshape(x, [-1, self.maxout_unit, self.fc_unit])
x = tf.reduce_max(x, reduction_indices=[1], name='d-reduce_max-1')
x = tf.nn.dropout(x, .5)
x = t.dense(x, 1, name='d-fc-2')
x = tf.nn.sigmoid(x)
return x
def encoder(self, x, reuse=None):
"""
:param x: Input images (32x32x3 or 64x64x3)
:param reuse: re-usable
:return: embeddings
"""
with tf.variable_scope('encoder', reuse=reuse):
repeat = int(np.log2(self.height)) - 2
x = t.conv2d(x, f=self.df_dim, name="enc-conv2d-1")
x = tf.nn.elu(x)
for i in range(1, repeat + 1):
f = self.df_dim * i
x = t.conv2d(x, f, 3, 1, name="enc-conv2d-%d" % (i * 2))
x = tf.nn.elu(x)
x = t.conv2d(x, f, 3, 1, name="enc-conv2d-%d" % (i * 2 + 1))
x = tf.nn.elu(x)
if i < repeat:
"""
You can choose one of them. max-pool or avg-pool or conv-pool.
Speed Order : conv-pool > avg-pool > max-pool. i guess :)
"""
x = t.conv2d(x, f, 3, 2, name='enc-conv2d-pool-%d' % i) # conv pooling
x = tf.nn.elu(x)
x = t.flatten(x)
x = t.dense(x, self.z_dim, name='enc-fc-1') # normally, (-1, 128)
return x
def generator(self, z, reuse=None):
"""
:param z: embeddings
:param reuse: re-usable
:return: logits
"""
with tf.variable_scope("generator", reuse=reuse):
repeat = int(np.log2(self.height)) - 2
x = t.dense(z, self.z_dim * 8 * 8, name='gen-fc-1')
x = tf.nn.elu(x)
x = tf.reshape(x, [-1, 8, 8, self.z_dim])
# shortcut = tf.identity(x, name='shortcut')
for i in range(1, repeat + 1):
x = t.conv2d(x, f=self.gf_dim, name="gen-conv2d-%d" % (i * 2 - 1))
x = tf.nn.elu(x)
x = t.conv2d(x, f=self.gf_dim, name="gen-conv2d-%d" % (i * 2))
x = tf.nn.elu(x)
# if i < 3:
# x = tf.add(x, shortcut)
if i < repeat:
x = t.up_sampling(x, tf.image.ResizeMethod.NEAREST_NEIGHBOR) # NN up-sampling
x = t.conv2d(x, f=self.channel, name='gen-conv2d-%d' % (2 * repeat + 1))
x = tf.nn.tanh(x)
return x
def discriminator(self, x, reuse=None):
with tf.variable_scope('discriminator', reuse=reuse):
x = self.residual_block_init(x,
self.z_dim,
reuse=reuse,
name='disc-res_block_init')
x = self.residual_block(x,
self.z_dim,
reuse=reuse,
sampling='down',
name='disc-res_block-1')
x = self.residual_block(x,
self.z_dim,
reuse=reuse,
sampling='none',
name='disc-res_block-2')
x = self.residual_block(x,
self.z_dim,
reuse=reuse,
sampling='none',
name='disc-res_block-3')
x = tf.nn.relu(x)
x = t.global_avg_pooling(x)
x = t.dense(x, 1, name='disc-fc-1')
return x
def generator(self, z, reuse=None):
"""
:param z: embeddings
:param reuse: re-usable
:return: logits
"""
with tf.variable_scope("generator", reuse=reuse):
repeat = int(np.log2(self.height)) - 2
x = t.dense(z, self.z_dim * 8 * 8, name='g-fc-1')
x = tf.nn.elu(x)
x = tf.reshape(x, [-1, 8, 8, self.z_dim])
for i in range(1, repeat + 1):
x = t.conv2d(x, f=self.gf_dim, name="g-conv-%d" % (i * 2 - 1))
x = tf.nn.elu(x)
x = t.conv2d(x, f=self.gf_dim, name="g-conv-%d" % (i * 2))
x = tf.nn.elu(x)
if i < repeat:
x = t.resize_nn(x,
x.get_shape().as_list()[1] *
2) # NN up-sampling
x = t.conv2d(x, f=self.channel)
return x
def generator(self, z, y, reuse=None, is_train=True):
"""
# Following a G Network, CiFar-like-hood, referred in the paper
:param z: noise
:param y: image label
:param reuse: re-usable
:param is_train: trainable
:return: prob
"""
with tf.variable_scope("generator", reuse=reuse):
x = tf.concat([z, y], axis=1) # (-1, 110)
x = t.dense(x, self.gf_dim, name='gen-fc-1')
x = tf.nn.relu(x)
x = tf.reshape(x, (-1, 4, 4, 24))
for i in range(1, 3):
x = t.deconv2d(x,
self.gf_dim // (2**i),
5,
2,
name='gen-deconv2d-%d' % (i + 1))
x = t.batch_norm(x,
is_train=is_train,
reuse=reuse,
name="gen-bn-%d" % i)
x = tf.nn.relu(x)
x = t.deconv2d(x, self.channel, 5, 2, name='gen-deconv2d-4')
x = tf.nn.tanh(x) # scaling to [-1, 1]
return x
def generator(self, z, y=None, reuse=None, is_train=True):
"""
:param z: embeddings
:param y: labels
:param reuse: re-usable
:param is_train: en/disable batch_norm, default True
:return: prob
"""
with tf.variable_scope("generator", reuse=reuse):
if y:
raise NotImplemented("[-] Not Implemented Yet...")
x = t.dense(z, f=self.fc_unit, name='gen-fc-0')
x = tf.nn.leaky_relu(x)
x = tf.reshape(x, [-1, 8, 8, self.fc_unit // (8 * 8)])
for i in range(1, 4):
x = t.deconv2d(x, f=self.gf_dim * (2 ** i), name="gen-conv2d-%d" % i)
x = t.batch_norm(x, is_train=is_train)
x = tf.nn.leaky_relu(x)
x = t.deconv2d(x, f=3, s=1, name="gen-conv2d-4") # (-1, 64, 64, 3)
x = tf.sigmoid(x) # [0, 1]
return x
def discriminator(self, x, y=None, reuse=None, is_train=True):
"""
:param x: images
:param y: labels
:param reuse: re-usable
:param is_train: en/disable batch_norm, default True
:return: logits
"""
with tf.variable_scope("discriminator", reuse=reuse):
if y:
raise NotImplemented("[-] Not Implemented Yet...")
x = t.conv2d(x, f=self.gf_dim * 1, name="disc-conv2d-0")
x = tf.nn.leaky_relu(x)
for i in range(1, 4):
x = t.conv2d(x, f=self.gf_dim * (2 ** i), name="disc-conv2d-%d" % i)
x = t.batch_norm(x, is_train=is_train)
x = tf.nn.leaky_relu(x)
feature_match = x # (-1, 8, 8, 512)
x = tf.layers.flatten(x)
x = t.dense(x, 1, name='disc-fc-0')
return feature_match, x
def discriminator(self, x, reuse=None):
"""
# Following a D Network, CiFar-like-hood, referred in the paper
:param x: image, shape=(-1, 28, 28, 1)
:param reuse: re-usable
:return: logits, networks
"""
with tf.variable_scope("discriminator", reuse=reuse):
x = t.conv2d(x, self.df_dim, k=3, s=2, name='d-conv-0')
x = tf.nn.leaky_relu(x)
x = tf.layers.dropout(x, 0.5, name='d-dropout-0')
for i in range(1, 2 * 2 + 1):
f = self.df_dim * (i + 1)
x = t.conv2d(x, f=f, k=3, s=(i % 2 + 1), name='d-conv-%d' % i)
x = t.batch_norm(x)
x = tf.nn.leaky_relu(x)
x = tf.layers.dropout(x, 0.5, name='d-dropout-%d' % i)
x = tf.layers.flatten(x)
x = t.dense(x, self.fc_unit * 2, name='d-fc-1')
net = tf.nn.leaky_relu(x)
x = tf.layers.dense(net, 1, name='d-fc-2') # logits
return x, net
def generator(self, z, reuse=None, is_train=True):
"""
# referred architecture in the paper
: (1024)4c - (512)4c2s - (256)4c2s - (128)4c2s - (3)4c2s
:param z: embeddings
:param reuse: re-usable
:param is_train: trainable
:return: prob
"""
with tf.variable_scope("generator", reuse=reuse):
x = t.dense(z, self.gf_dim * 8 * 4 * 4, name='gen-fc-1')
x = tf.reshape(x, (-1, 4, 4, self.gf_dim * 8))
x = t.batch_norm(x, is_train=is_train, name='gen-bn-1')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.gf_dim * 4, 4, 2, name='gen-deconv2d-1')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-2')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.gf_dim * 2, 4, 2, name='gen-deconv2d-2')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-3')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.gf_dim * 1, 4, 2, name='gen-deconv2d-3')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-4')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.channel, 4, 2, name='gen-deconv2d-4')
x = tf.nn.tanh(x)
return x
def generator(self, z, c, reuse=None, is_train=True):
"""
:param z: 139 z-noise
:param c: 10 categories * 10 dimensions
:param reuse: re-usable
:param is_train: trainable
:return: prob
"""
with tf.variable_scope("generator", reuse=reuse):
x = tf.concat([z, c], axis=1) # (-1, 128 + 1 + 10)
x = t.dense(x, 2 * 2 * 512, name='gen-fc-1')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-1')
x = tf.nn.relu(x)
x = tf.reshape(x, (-1, 2, 2, 512))
x = t.deconv2d(x, self.gf_dim * 8, 4, 2, name='gen-deconv2d-1')
x = t.batch_norm(x, is_train=is_train, name='gen-bn-2')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.gf_dim * 4, 4, 2, name='gen-deconv2d-2')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.gf_dim * 2, 4, 2, name='gen-deconv2d-3')
x = tf.nn.relu(x)
x = t.deconv2d(x, self.gf_dim * 1, 4, 2, name='gen-deconv2d-4')
x = tf.nn.relu(x)
x = t.deconv2d(x, 3, 4, 2, name='gen-deconv2d-5')
x = tf.nn.tanh(x)
return x
def generator(self, x, y, z, scale=32, reuse=None, do_rate=0.5):
"""
:param x: images to fake
:param y: classes
:param z: noise
:param scale: image size
:param reuse: variable re-use
:param do_rate: dropout rate
:return: logits
"""
assert (scale % 8 == 0) # 32, 16, 8
with tf.variable_scope('generator_{0}'.format(scale), reuse=reuse):
if scale == 8:
h = tf.concat([z, y], axis=1)
h = t.dense(h, self.g_fc_unit, name='gen-fc-1')
h = tf.nn.relu(h)
h = tf.layers.dropout(h, do_rate, name='gen-dropout-1')
h = t.dense(h, self.g_fc_unit, name='gen-fc-2')
h = tf.nn.relu(h)
h = tf.layers.dropout(h, do_rate, name='gen-dropout-2')
h = t.dense(h, self.channel * 8 * 8, name='gen-fc-3')
h = tf.reshape(h, [-1, 8, 8, self.channel])
else:
y = t.dense(y, scale * scale, name='gen-fc-y')
y = tf.reshape(y, [-1, scale, scale, 1])
z = tf.reshape(z, [-1, scale, scale, 1])
h = tf.concat([z, y, x], axis=3) # concat into 5 dims
h = t.conv2d(h, self.gf_dim * 1, 5, 1, name='gen-deconv2d-1')
h = tf.nn.relu(h)
h = t.conv2d(h, self.gf_dim * 1, 5, 1, name='gen-deconv2d-2')
h = tf.nn.relu(h)
h = t.conv2d(h, self.channel, 5, 1, name='gen-conv2d-3')
h = tf.nn.tanh(h)
return h
def generator(self, z, y=None, share_params=False, reuse=False, name=""):
x = t.dense(z, self.fc_g_unit, reuse=share_params, name='gen-fc-1')
x = t.batch_norm(x, reuse=share_params, name='gen-bn-1')
x = t.prelu(x, reuse=share_params, name='gen-prelu-1')
x = t.dense(x,
self.gf_dim * 8 * 7 * 7,
reuse=share_params,
name='gen-fc-2')
x = t.batch_norm(x, reuse=share_params, name='gen-bn-2')
x = t.prelu(x, reuse=share_params, name='gen-prelu-2')
x = tf.reshape(x, (-1, 7, 7, self.gf_dim * 8))
for i in range(1, 3):
x = t.deconv2d(x,
f=self.gf_dim * 4 // i,
k=3,
s=2,
reuse=share_params,
name='gen-deconv2d-%d' % i)
x = t.batch_norm(x, reuse=share_params, name="gen-bn-%d" % (i + 2))
x = t.prelu(x, reuse=share_params, name='gen-prelu-%d' % (i + 2))
"""
x = z # tf.concat([z, y], axis=1)
loop = 5
for i in range(1, loop):
x = t.dense(x, self.fc_g_unit, reuse=share_params, name='gen-fc-%d' % i)
x = t.batch_norm(x, reuse=share_params, name='gen-bn-%d' % i)
x = t.prelu(x, reuse=share_params, name='gen-prelu-%d' % i)
"""
with tf.variable_scope("generator-%s" % name, reuse=reuse):
x = t.deconv2d(x,
f=self.channel,
k=6,
s=1,
reuse=False,
name='gen-' + name + '-deconv2d-3')
x = tf.nn.sigmoid(x, name='gen' + name + '-sigmoid-1')
"""
x = t.dense(x, self.n_input, reuse=False, name='gen-' + name + '-fc-%d' % loop)
x = tf.nn.sigmoid(x)
"""
return x
def discriminator(self, x, scope_name, reuse=None):
with tf.variable_scope("%s" % scope_name, reuse=reuse):
x = t.conv2d(x, f=self.df_dim, k=4, s=1) # 64 x 64 x 3
x = tf.nn.leaky_relu(x)
for i in range(np.log2(x.get_shape()[1]) - 2): # 0 ~ 3
x = t.conv2d(x, self.df_dim * (2 ** (i + 1)), k=4, s=2)
x = t.batch_norm(x)
x = tf.nn.leaky_relu(x)
# (-1, 4, 4, 512)
x = tf.layers.flatten(x)
x = t.dense(x, 512)
x = tf.nn.leaky_relu(x)
x = t.dense(x, 1)
x = tf.sigmoid(x)
return x
def discriminator(self, x, reuse=None):
with tf.variable_scope('discriminator', reuse=reuse):
for i in range(1, 4):
x = t.conv2d(x, self.gf_dim * (2 ** (i - 1)), 3, 2, name='disc-conv2d-%d' % i)
x = t.batch_norm(x, name='disc-bn-%d' % i)
x = tf.nn.leaky_relu(x, alpha=0.3)
x = tf.layers.flatten(x)
x = t.dense(x, 1, name='disc-fc-1')
return x
def generator(self, z, reuse=None, is_train=True):
with tf.variable_scope('generator', reuse=reuse):
x = t.dense(z, self.fc_unit, name='gen-fc-0')
x = t.batch_norm(x, is_train=is_train)
x = tf.nn.leaky_relu(x)
x = t.dense(x, self.gf_dim * 4 * 7 * 7, name='gen-fc-1')
x = t.batch_norm(x, is_train=is_train)
x = tf.nn.leaky_relu(x)
x = tf.reshape(x, [-1, 7, 7, self.gf_dim * 4])
x = t.deconv2d(x, self.gf_dim * 2, name='gen-deconv2d-0')
x = t.batch_norm(x, is_train=is_train)
x = tf.nn.leaky_relu(x)
logits = t.deconv2d(x, self.channel, name='gen-deconv2d-1')
prob = tf.nn.sigmoid(logits)
return prob
