In keeping with the theme of autoencoders, in this notebook I perform an experiment with an autoencoder to resolve a problem I had in training a GAN.