Public Code Repository

The workshops and seminars resulted in many Jupyter notebooks illustrating foundational models and datasets in current AI. The code is written in PyTorch or Jax. Click on the button below to access the code repository.

One of the Jupyter Notebooks available illustartes the Proximal Policy Optimization algorithm for Reinforcement Learning. Below is a gif of the Cartpole environment which solved using PPO.

Another Jupyter Notebook illustrates the Variational Autoencoder algorithm for unsupervised learning. The numbers in the image below were diffused into Gaussian noise and then regenerated by VAE.

We also explored Denoising Diffusion models. The cats in the images below were diffused into Gaussian noise (and then regenerated).