def discriminator(dis_inputs, is_training=True):
# Define the discriminator block
def discriminator_block(inputs, output_channel, kernel_size, stride,
scope):
with tf.variable_scope(scope):
net = conv2(inputs,
kernel_size,
output_channel,
stride,
use_bias=False,
scope='conv1')
# net = batchnorm(net, is_training)
net = lrelu(net, 0.2)
return net
with tf.variable_scope('discriminator_unit'):
# The input layer
with tf.variable_scope('input_stage'):
net = conv2(dis_inputs, 3, 64, 1, scope='conv')
net = lrelu(net, 0.2)
# The discriminator block part
# block 1
net = discriminator_block(net, 64, 3, 2, 'disblock_1')
# block 2
net = discriminator_block(net, 128, 3, 1, 'disblock_2')
# block 3
net = discriminator_block(net, 128, 3, 2, 'disblock_3')
# block 4
net = discriminator_block(net, 256, 3, 1, 'disblock_4')
# block 5
net = discriminator_block(net, 256, 3, 2, 'disblock_5')
# block 6
net = discriminator_block(net, 512, 3, 1, 'disblock_6')
# block_7
net = discriminator_block(net, 512, 3, 2, 'disblock_7')
# The dense layer 1
with tf.variable_scope('dense_layer_1'):
net = slim.flatten(net)
net = denselayer(net, 1024)
net = lrelu(net, 0.2)
# The dense layer 2
with tf.variable_scope('dense_layer_2'):
logits = denselayer(net, 1)
prob = tf.nn.sigmoid(logits)
return logits, prob
def discriminatorDense(dis_inputs, is_training=True):
# Define the discriminator block
def discriminator_block(inputs, output_channel, kernel_size, stride, scope):
with tf.variable_scope(scope):
net = conv2(inputs, kernel_size, output_channel, stride, use_bias=False, scope='conv1')
# net = batchnorm(net, is_training)
net = lrelu(net, 0.2)
return net
with tf.variable_scope('discriminator_unit'):
# The input layer
with tf.variable_scope('input_stage'):
net = conv2(dis_inputs, 3, 64, 1, scope='conv')
net = lrelu(net, 0.2)
# The discriminator block part
# block 1
net = discriminator_block(net, 64, 3, 2, 'disblock_1')
# block 2
net = discriminator_block(net, 64, 3, 2, 'disblock_2')
# block 3
net = discriminator_block(net, 64, 3, 1, 'disblock_3')
# The dense block part
# Define the denseblock configuration
layer_per_block = 8
bottleneck_scale = 4
growth_rate = 12
transition_output_channel = 128
with tf.variable_scope('denseBlock_1'):
net = denseBlock(net, layer_per_block, bottleneck_scale, growth_rate, is_training)
with tf.variable_scope('transition_layer_1'):
net = transitionLayer(net, transition_output_channel, is_training)
# The dense layer 1
with tf.variable_scope('dense_layer_1'):
net = slim.flatten(net)
net = denselayer(net, 1024)
net = lrelu(net, 0.2)
# The dense layer 2
with tf.variable_scope('dense_layer_2'):
net = denselayer(net, 1)
net = tf.nn.sigmoid(net)
return net
def discriminator_emb(dis_embeddings, is_training=True):
# Define the discriminator block
def discriminator_block(inputs, output_channel, scope):
with tf.variable_scope(scope):
net = denselayer(inputs, output_channel)
net = lrelu(net, 0.2)
return net
with tf.variable_scope('discriminator_unit'):
# The discriminator block part
# block 1
net = discriminator_block(dis_embeddings, 1024, 'disblock_1')
# block 2
net = discriminator_block(net, 1024, 'disblock_2')
# block 3
net = discriminator_block(net, 512, 'disblock_3')
# block 4
net = discriminator_block(net, 512, 'disblock_4')
# The dense layer 1
with tf.variable_scope('dense_layer_1'):
net = denselayer(net, 128)
net = lrelu(net, 0.2)
# The dense layer 2
with tf.variable_scope('dense_layer_2'):
net = denselayer(net, 1)
net = tf.nn.sigmoid(net)
return net
def fnet(fnet_input, reuse=False):
def down_block(inputs, output_channel=64, stride=1, scope='down_block'):
with tf.compat.v1.variable_scope(scope):
net = conv2(inputs,
3,
output_channel,
stride,
use_bias=True,
scope='conv_1')
net = lrelu(net, 0.2)
net = conv2(net,
3,
output_channel,
stride,
use_bias=True,
scope='conv_2')
net = lrelu(net, 0.2)
net = maxpool(net)
return net
def up_block(inputs, output_channel=64, stride=1, scope='up_block'):
with tf.compat.v1.variable_scope(scope):
net = conv2(inputs,
3,
output_channel,
stride,
use_bias=True,
scope='conv_1')
net = lrelu(net, 0.2)
net = conv2(net,
3,
output_channel,
stride,
use_bias=True,
scope='conv_2')
net = lrelu(net, 0.2)
new_shape = tf.shape(net)[1:-1] * 2
net = tf.image.resize(net, new_shape)
return net
with tf.compat.v1.variable_scope('autoencode_unit', reuse=reuse):
net = down_block(fnet_input, 32, scope='encoder_1')
net = down_block(net, 64, scope='encoder_2')
net = down_block(net, 128, scope='encoder_3')
net = up_block(net, 256, scope='decoder_1')
net = up_block(net, 128, scope='decoder_2')
net1 = up_block(net, 64, scope='decoder_3')
with tf.compat.v1.variable_scope('output_stage'):
net = conv2(net1, 3, 32, 1, scope='conv1')
net = lrelu(net, 0.2)
net2 = conv2(net, 3, 2, 1, scope='conv2')
net = tf.tanh(net2) * 24.0
# the 24.0 is the max Velocity, details can be found in TecoGAN paper
return net
def down_block(inputs, output_channel=64, stride=1, scope='down_block'):
with tf.compat.v1.variable_scope(scope):
net = conv2(inputs,
3,
output_channel,
stride,
use_bias=True,
scope='conv_1')
net = lrelu(net, 0.2)
net = conv2(net,
3,
output_channel,
stride,
use_bias=True,
scope='conv_2')
net = lrelu(net, 0.2)
net = maxpool(net)
return net
def up_block(inputs, output_channel=64, stride=1, scope='up_block'):
with tf.compat.v1.variable_scope(scope):
net = conv2(inputs,
3,
output_channel,
stride,
use_bias=True,
scope='conv_1')
net = lrelu(net, 0.2)
net = conv2(net,
3,
output_channel,
stride,
use_bias=True,
scope='conv_2')
net = lrelu(net, 0.2)
new_shape = tf.shape(net)[1:-1] * 2
net = tf.image.resize(net, new_shape)
return net
def discriminator_block(inputs, output_channel, kernel_size, stride,
scope):
with tf.variable_scope(scope):
net = conv2(inputs,
kernel_size,
output_channel,
stride,
use_bias=False,
scope='conv1')
# net = batchnorm(net, is_training)
net = lrelu(net, 0.2)
return net
def discriminator_block(inputs, output_channel, kernel_size, stride,
scope):
with tf.variable_scope(scope):
net = ops.conv3d(inputs,
kernel_size,
output_channel,
stride,
use_bias=False,
scope='conv1')
if (FLAGS.GAN_type == 'GAN'):
net = ops.batchnorm(net, FLAGS.is_training)
net = ops.lrelu(net, 0.2)
return net
def discriminator_feature(dis_features, is_training=True):
# Define the discriminator block
def discriminator_block(inputs, output_channel, kernel_size, stride,
scope):
with tf.variable_scope(scope):
net = conv2(inputs,
kernel_size,
output_channel,
stride,
use_bias=False,
scope='conv1')
# net = batchnorm(net, is_training)
net = lrelu(net, 0.2)
return net
with tf.variable_scope('discriminator_unit'):
# The discriminator block part
# block 1
net = discriminator_block(dis_features, 256, 3, 1, 'disblock_1')
# block 2
net = discriminator_block(net, 256, 3, 2, 'disblock_2')
# block 3
net = discriminator_block(net, 512, 3, 1, 'disblock_3')
# block 4
net = discriminator_block(net, 512, 3, 2, 'disblock_4')
# The dense layer 1
with tf.variable_scope('dense_layer_1'):
net = slim.flatten(net)
net = denselayer(net, 1024)
net = lrelu(net, 0.2)
# The dense layer 2
with tf.variable_scope('dense_layer_2'):
net = denselayer(net, 1)
net = tf.nn.sigmoid(net)
return net
def discriminator(dis_inputs, FLAGS=None):
if FLAGS is None:
raise ValueError('No FLAGS is provided for discriminator')
# Define the discriminator block
def discriminator_block(inputs, output_channel, kernel_size, stride,
scope):
with tf.variable_scope(scope):
net = ops.conv3d(inputs,
kernel_size,
output_channel,
stride,
use_bias=False,
scope='conv1')
if (FLAGS.GAN_type == 'GAN'):
net = ops.batchnorm(net, FLAGS.is_training)
net = ops.lrelu(net, 0.2)
return net
# with tf.device('/gpu:0'):
with tf.variable_scope('discriminator_unit'):
# The input layer
with tf.variable_scope('input_stage'):
net = ops.conv3d(dis_inputs, 3, 64, 1, scope='conv')
net = ops.lrelu(net, 0.2) # (m, 64, 64, 64, 64)
# The discriminator block part
# block 1
net = discriminator_block(net, 64, 3, 2,
'disblock_1') # (m, 32, 32, 32, 64)
# block 2
net = discriminator_block(net, 128, 3, 1,
'disblock_2') # (m, 32, 32, 32, 128)
# block 3
net = discriminator_block(net, 128, 3, 2,
'disblock_3') # (m, 16, 16, 16, 128)
# block 4
net = discriminator_block(net, 256, 3, 1,
'disblock_4') # (m, 16, 16, 16, 256)
# block 5
net = discriminator_block(net, 256, 3, 2,
'disblock_5') # (m, 8, 8, 8, 256)
# block 6
net = discriminator_block(net, 512, 3, 1,
'disblock_6') # (m, 8, 8, 8, 512)
# block_7
net = discriminator_block(net, 512, 3, 2,
'disblock_7') # (m, 4, 4, 4, 512)
# block_8
net = discriminator_block(net, 1024, 3, 1,
'disblock_8') # (m, 4, 4, 4, 1024)
# block_9
net = discriminator_block(net, 1024, 3, 2,
'disblock_9') # (m, 2, 2, 2, 1024)
# The dense layer 1
with tf.variable_scope('dense_layer_1'):
net = tf.contrib.layers.flatten(net)
net = ops.denselayer(net, 1024)
net = ops.lrelu(net, 0.2)
# The dense layer 2
with tf.variable_scope('dense_layer_2'):
net = ops.denselayer(net, 1)
net = tf.nn.sigmoid(net)
return net
def discriminator_block(inputs, output_channel, scope):
with tf.variable_scope(scope):
net = denselayer(inputs, output_channel)
net = lrelu(net, 0.2)
return net
