generative-models

Why we included this project

Reading a paper on GANs is one thing; seeing the same model written out in clean, runnable code is another. This repository collects working implementations of the major generative model families in PyTorch and TensorFlow. The GAN lineup alone spans the classics, from vanilla and Wasserstein to InfoGAN, DiscoGAN, and boundary equilibrium, and it continues with variational autoencoders, restricted Boltzmann machines, and a Helmholtz machine. Each model comes with training code and saves generated samples to an output folder, so you can run an experiment and inspect what it produces. That makes it a solid study companion for a student learning the mechanics or a researcher revisiting a classic paper, and a handy way to understand a generative architecture before picking a production tool. The code is meant for learning, not deployment, so expect to adapt it rather than run it as-is.