Get up and running with Llama 3.3, DeepSeek-R1, Phi-4, Gemma 2, and other large language models.

CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image

OpenVLA: An open-source vision-language-action model for robotic manipulation.

Official repository for "LIV: Language-Image Representations and Rewards for Robotic Control" (ICML 2023)

[ICML 2024] The offical Implementation of "DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning"

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

RL implementations

Official Repository for "Eureka: Human-Level Reward Design via Coding Large Language Models" (ICLR 2024)

A Zsh theme

Isaac Gym Reinforcement Learning Environments

High-quality single-file implementations of SOTA Offline and Offline-to-Online RL algorithms: AWAC, BC, CQL, DT, EDAC, IQL, SAC-N, TD3+BC, LB-SAC, SPOT, Cal-QL, ReBRAC

[NeurIPS'21 Outstanding Paper] Library for reliable evaluation on RL and ML benchmarks, even with only a handful of seeds.

High-quality single file implementation of Deep Reinforcement Learning algorithms with research-friendly features (PPO, DQN, C51, DDPG, TD3, SAC, PPG)

PyTorch version of Stable Baselines, reliable implementations of reinforcement learning algorithms.

Tuning LLMs with no tears💦; Sample Design Engineering (SDE) for more efficient downstream-tuning.

The official Meta Llama 3 GitHub site

Example models using DeepSpeed

Code for Paper (ReMax: A Simple, Efficient and Effective Reinforcement Learning Method for Aligning Large Language Models)

An open source implementation of CLIP.

Code for RoboFlamingo

SMAC: The StarCraft Multi-Agent Challenge

An open-source framework for training large multimodal models.

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

Train transformer language models with reinforcement learning.

Official Code for RVT-2 and RVT

Code for the paper "3D Diffuser Actor: Policy Diffusion with 3D Scene Representations"

CALVIN - A benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks