GAN+Encoder with ASCII Image Visualization

This video shows how the print_ascii_image function from the Datawaza library can be used to visualize the output of a generative model. In this example, it's a Generative Adversarial Network (GAN) that has a Generator and Discriminator. In addition, I added an Encoder that encodes x_real to generate the z latent variables. These are then run through the Generator to evaluate how well the GAN is performing (not random z). This makes it easier to see how well the GAN model is performing, because there's at least a similar type of image shown adjacent to it. The training of the GAN is completely separate from the Encoder (it doesn't know it exists). The print_ascii_image function can be found in the Datawaza library: https://github.com/jbeno/datawaza https://www.datawaza.com/en/latest/userguide.html#Print-ASCII-Image https://www.datawaza.com/en/latest/explore.html#datawaza.explore.print_ascii_image