dcgan-2.py

noise_shape = 100

#  Generator will upsample our seed using convolutional transpose layers (upsampling layers)
def generator_model():
  generator=Sequential()
  
  # Random noise to 4x4x512 image
  generator.add(Dense(4*4*512, input_shape=[noise_shape]))
  
  #  Next, add a reshape layer to the network to reshape the tensor from the 
  # last layer to a tensor of a shape of (4, 4, 512):
  generator.add(Reshape([4,4,512]))
  generator.add(Conv2DTranspose(256, kernel_size=4, strides=2, padding="same"))
  # BatchNormalization is added to the model after the hidden layer, but before the activation, such as LeakyReLU.
  generator.add(BatchNormalization())
  generator.add(LeakyReLU(alpha=0.2))
  
  generator.add(Conv2DTranspose(128, kernel_size=4, strides=2, padding="same"))
  generator.add(LeakyReLU(alpha=0.2))
  
  generator.add(BatchNormalization())
  generator.add(Conv2DTranspose(64, kernel_size=4, strides=2, padding="same"))
  generator.add(LeakyReLU(alpha=0.2))
  generator.add(BatchNormalization())
  generator.add(Conv2DTranspose(3, kernel_size=4, strides=2, padding="same",
                                  activation='sigmoid'))
  return generator

generator = generator_model()
generator.summary()


def discriminator_model():
  discriminator = Sequential()
  discriminator.add(Conv2D(64, (3,3), strides=(2, 2), padding='same', input_shape=[64,64, 3]))
  discriminator.add(LeakyReLU(alpha=0.2))
  discriminator.add(Dropout(0.4))
  discriminator.add(Conv2D(64, (3,3), strides=(2, 2), padding='same'))
  discriminator.add(BatchNormalization())
  discriminator.add(LeakyReLU(alpha=0.2))  
  discriminator.add(Dropout(0.4))
  discriminator.add(Flatten())
  discriminator.add(Dense(1, activation='sigmoid'))
  return discriminator

discriminator = discriminator_model()

discriminator.summary()