def discriminator(self, x, reuse=None):
"""
:param x: images
:param y: labels
:param reuse: re-usable
:return: classification, probability (fake or real), network
"""
with tf.variable_scope("discriminator", reuse=reuse):
f = self.gf_dim
x = t.conv2d_alt(x, f, 4, 2, pad=1, sn=True, name='disc-conv2d-1')
x = tf.nn.leaky_relu(x, alpha=0.1)
for i in range(self.n_layer // 2):
x = t.conv2d_alt(x, f * 2, 4, 2, pad=1, sn=True, name='disc-conv2d-%d' % (i + 2))
x = tf.nn.leaky_relu(x, alpha=0.1)
f *= 2
# Self-Attention Layer
x = self.attention(x, f, reuse=reuse)
for i in range(self.n_layer // 2, self.n_layer):
x = t.conv2d_alt(x, f * 2, 4, 2, pad=1, sn=True, name='disc-conv2d-%d' % (i + 2))
x = tf.nn.leaky_relu(x, alpha=0.1)
f *= 2
x = t.flatten(x)
x = t.dense_alt(x, 1, sn=True, name='disc-fc-1')
return x
def discriminator(self, x, reuse=None):
"""
:param x: images
:param reuse: re-usable
:return: classification, probability (fake or real), network
"""
with tf.variable_scope("discriminator", reuse=reuse):
f = 1 * self.channel
x = self.res_block(x, f=f, scale_type="down", name="disc-res1")
x = self.self_attention(x, f_=f)
for i in range(4):
f *= 2
x = self.res_block(x, f=f, scale_type="down", name="disc-res%d" % (i + 1))
x = self.res_block(x, f=f, scale_type="down", use_bn=False, name="disc-res5")
x = tf.nn.relu(x)
with tf.name_scope("global_sum_pooling"):
x_shape = x.get_shape().as_list()
x = tf.reduce_mean(x, axis=-1) * (x_shape[1] * x_shape[2])
x = t.dense_alt(x, 1, sn=True, name='disc-dense-last')
return x
def generator(self, z, reuse=None, is_train=True):
"""
:param z: noise
:param y: image label
:param reuse: re-usable
:param is_train: trainable
:return: prob
"""
with tf.variable_scope("generator", reuse=reuse):
f = self.gf_dim * 8
x = t.dense_alt(z, 4 * 4 * f, sn=True, name='gen-fc-1')
x = tf.reshape(x, (-1, 4, 4, f))
for i in range(self.n_layer // 2):
if self.up_sampling:
x = t.up_sampling(x, interp=tf.image.ResizeMethod.NEAREST_NEIGHBOR)
x = t.conv2d_alt(x, f // 2, 5, 1, pad=2, sn=True, use_bias=False, name='gen-conv2d-%d' % (i + 1))
else:
x = t.deconv2d_alt(x, f // 2, 4, 2, sn=True, use_bias=False, name='gen-deconv2d-%d' % (i + 1))
x = t.batch_norm(x, is_train=is_train, name='gen-bn-%d' % i)
x = tf.nn.relu(x)
f //= 2
# Self-Attention Layer
x = self.attention(x, f, reuse=reuse)
for i in range(self.n_layer // 2, self.n_layer):
if self.up_sampling:
x = t.up_sampling(x, interp=tf.image.ResizeMethod.NEAREST_NEIGHBOR)
x = t.conv2d_alt(x, f // 2, 5, 1, pad=2, sn=True, use_bias=False, name='gen-conv2d-%d' % (i + 1))
else:
x = t.deconv2d_alt(x, f // 2, 4, 2, sn=True, use_bias=False, name='gen-deconv2d-%d' % (i + 1))
x = t.batch_norm(x, is_train=is_train, name='gen-bn-%d' % i)
x = tf.nn.relu(x)
f //= 2
x = t.conv2d_alt(x, self.channel, 5, 1, pad=2, sn=True, name='gen-conv2d-%d' % (self.n_layer + 1))
x = tf.nn.tanh(x)
return x
def generator(self, z, reuse=None):
"""
:param z: noise
:param reuse: re-usable
:return: prob
"""
with tf.variable_scope("generator", reuse=reuse):
# split
z = tf.split(z, num_or_size_splits=4, axis=-1) # expected [None, 32] * 4
# linear projection
x = t.dense_alt(z, f=4 * 4 * 16 * self.channel, sn=True, use_bias=False, name="gen-dense-1")
x = tf.nn.relu(x)
x = tf.reshape(x, (-1, 4, 4, 16 * self.channel))
res = x
f = 16 * self.channel
for i in range(4):
res = self.res_block(res,
f=f,
scale_type="up",
name="gen-res%d" % (i + 1))
f //= 2
x = self.self_attention(res, f_=f * 2)
x = self.res_block(x, f=1 * self.channel, scale_type="up", name="gen-res4")
x = t.batch_norm(x, name="gen-bn-last") # <- noise
x = tf.nn.relu(x)
x = t.conv2d_alt(x, f=self.channel, k=3, sn=True, name="gen-conv2d-last")
x = tf.nn.tanh(x)
return x