GANs from scratch.

A gentle explanation/discussion on Generative Adversarial networks along with implementation in PyTorch

Generative adversarial networks(GANs) took AI by storm last year with those impressive human-like faces. They were really cool weren’t they? They are basically generated from nothing.

Nothing??. *Coughs* Show-off *Coughs*. You people used data to train the model. We know how these “machine learning” things work. It's just an Input-Output function approximation. There is nothing impressive about it…