Official codebase used to develop Vision Transformer, SigLIP, MLP-Mixer, LiT and more.

A Production-ready Reinforcement Learning AI Agent Library brought by the Applied Reinforcement Learning team at Meta.

A suite of image and video neural tokenizers

Evaluating and reproducing real-world robot manipulation policies (e.g., RT-1, RT-1-X, Octo) in simulation under common setups (e.g., Google Robot, WidowX+Bridge) (CoRL 2024)

Build your own visual reasoning model

Cosmos-Predict1 is a collection of general-purpose world foundation models for Physical AI that can be fine-tuned into customized world models for downstream applications.

Representation Learning MSc course Summer Semester 2023

Code for reproducing our paper "Are Sparse Autoencoders Useful? A Case Study in Sparse Probing"

Coherent Soft Imitation Learning