Example of a simple GAN model using keras.
The generator is trained against the discriminator using the script train_generator.py.
The generator learns to create 28x28 images of clothes and shoes similar to the ones in the Fashion-MNIST dataset.
A saved generator model can be tested using the script test_generator.py. It will output a matrix of mxn results obtained feeding the generator network with mxn random vectors obtained from a normal distribution.
Example of a random set of 4x4 images obtained from the network:
