SLIP: Self-supervision meets Language-Image Pre-training What you can find in this repo: Pre-trained models (with ViT-Small, Base, Large) and code to