Simple GAN with Keras

Note

You can follow along with the code in the Jupyter notebook ch-14a_SimpleGAN.

Now let us implement the same model in Keras: 

1. The hyper-parameter definitions remain the same as the last section:

# graph hyperparameters
g_learning_rate = 0.00001
d_learning_rate = 0.01
n_x = 784  # number of pixels in the MNIST image 
# number of hidden layers for generator and discriminator
g_n_layers = 3
d_n_layers = 1
# neurons in each hidden layer
g_n_neurons = [256, 512, 1024]
d_n_neurons = [256]

2. Next, define the generator network:

# define generator

g_model = Sequential()
g_model.add(Dense(units=g_n_neurons[0], 
                  input_shape=(n_z,),
                  name='g_0'))
g_model.add(LeakyReLU())
for i in range(1,g_n_layers):
g_model.add(Dense(units=g_n_neurons[i],
                      name='g_{}'.format(i)
                     ))
    g_model.add(LeakyReLU())
g_model.add(Dense(units=n_x, activation='tanh',name='g_out'))
print('Generator:')
g_model.summary()
g_model.compile...