GPU-optimized version of the MuJoCo physics simulator, designed for NVIDIA hardware.

NVIDIA Isaac GR00T N1 is the world's first open foundation model for generalized humanoid robot reasoning and skills.

A minimal and stable PPO.

Code for SAPG: Split and Aggregate Policy Gradients (ICML 2024)

Long context evaluation for large language models

LLM training in simple, raw C/CUDA

LLMPerf is a library for validating and benchmarking LLMs

Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads

Open-Source Implementations of Multi-Modal Diffusion Models Optimized for Highest Quality and Ease of Use

RayLLM - LLMs on Ray (Archived). Read README for more info.

A collection of notes and implementations of machine learning algorithms from Andrew Ng's machine learning specialization.

Language model alignment-focused deep learning curriculum

High-speed download of LLaMA, Facebook's 65B parameter GPT model

Reinforcement learning algorithms in RLlib

EasyRL: An easy-to-use and comprehensive reinforcement learning package.

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

Multi-Joint dynamics with Contact. A general purpose physics simulator.

Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

Multi Task RL Baselines

Collections of robotics environments geared towards benchmarking multi-task and meta reinforcement learning

My adventures exploring soccer data analysis

A toolkit for reproducible reinforcement learning research.