Generative Adversarial Networks - GANs Intro

DCGAN Project - Tracking GAN Training Performance

We now have built our DCGAN that we'll train on MNIST images. Before we begin training, we'll introduce a new class that will allow us to keep track of useful training statistics during the training process and display the progress of each epoch.

We'll make use of this class to customize the training metrics we want to track and to easily visually judge performance of our GAN during training. Managing the training progress in a separate class allows us to keep the upcoming training loop code clutter-free.