Implementation of various machine learning related algorithms.
You like GAN ? We got GAN that generates images and GAN that learns to communicate secretly. Don't like the style of GANs ? We got style transfer. You want explainability ? We got RISE. Embeddings ? We got CLIP that gives you both image and text embedding (two for one!)! You prefer research on the architecture related questions ? I understand, we got forward forward and synthetic gradients. Scale ? We have studied scale like double decent.
For a more concrete list of implementations done you can see the readme for each specific year (I'll create a clean table in the future).