Advanced Deep Learning with Keras

Cross-Domain GANs
kernel_size=kernel_size,
strides=strides,
padding='same')
x = inputs
if instance_norm:
x = InstanceNormalization()(x)
if activation == 'relu':
x = Activation('relu')(x)
else:
x = LeakyReLU(alpha=0.2)(x)
x = conv(x)
x = concatenate([x, paired_inputs])
return x
Listing 7.1.2, cyclegan-7.1.1.py. Generator implementation in Keras:
def build_generator(input_shape,
output_shape=None,
kernel_size=3,
name=None):
"""The generator is a U-Network made of a 4-layer encoder
and a 4-layer decoder. Layer n-i is connected to layer i.
Arguments:
input_shape (tuple): input shape
output_shape (tuple): output shape
kernel_size (int): kernel size of encoder & decoder layers
name (string): name assigned to generator model
Returns:
generator (Model):
"""
inputs = Input(shape=input_shape)
channels = int(output_shape[-1])
e1 = encoder_layer(inputs,
32,
kernel_size=kernel_size,
activation='leaky_relu',
strides=1)
e2 = encoder_layer(e1,
64,
activation='leaky_relu',
[ 214 ]