Basic GAN

In this recipe, we look at the most basic GAN network for the Fashion-MNIST dataset.

Fashion-MNIST is a dataset of Zalando article images consisting of a training set of 60,000 examples; it also contains a test set of 10,000 examples. Each example is a 28 x 28 grayscale image, associated with a label from 10 classes.

Here are some example images from Fashion-MNIST:

Fashion MNIST is directly available in Keras.

Getting ready

Create a class called GAN. Import the relevant classes and initialize the variables:

from __future__ import print_function, division
from keras.datasets import fashion_mnist
from keras.layers import Input, Dense, Reshape, Flatten, Dropout
from keras.layers import BatchNormalization, Activation, ZeroPadding2D
from keras.layers.advanced_activations import LeakyReLU
from keras.layers.convolutional import UpSampling2D, Conv2D
from keras.models import Sequential, Model
from keras.optimizers import Adam
import matplotlib.pyplot as plt
import sys
import numpy as np

GAN...