🌾 OAT: A research-friendly framework for LLM online alignment, including reinforcement learning, preference learning, etc.

Defeating the Training-Inference Mismatch via FP16

Defeating the Training-Inference Mismatch via FP16

Unified framework for robot learning built on NVIDIA Isaac Sim

An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.

🤗 LeRobot: Making AI for Robotics more accessible with end-to-end learning

NVIDIA Isaac GR00T N1.5 - A Foundation Model for Generalist Robots.

Implement a ChatGPT-like LLM in PyTorch from scratch, step by step

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.

Foundation Model for Physics Simulations

Instant neural graphics primitives: lightning fast NeRF and more

The code for NeurIPS 2023 paper DSR

PyTorch code and models for VJEPA2 self-supervised learning from video.

Demo application for an event-driven multi-agent system for handling lead processing using an orchestrator built in Flink

Pytorch-based framework for solving parametric constrained optimization problems, physics-informed system identification, and parametric model predictive control.

Universal, autodiff-native software components for Simulation Intelligence. 📦

Geometric Neural Operators (GNPs) for machine learning tasks on point-cloud representations: curvature estimation, shape deformations, solvers for geometric PDEs. Also, includes weights of pretrain…

seqax = sequence modeling + JAX

[COLM 2025] Code for Paper: Learning Adaptive Parallel Reasoning with Language Models

EasyR1: An Efficient, Scalable, Multi-Modality RL Training Framework based on veRL

VIP cheatsheet for Stanford's CME 295 Transformers and Large Language Models

A simple, performant and scalable Jax LLM!

Ongoing research training transformer models at scale

The official Python SDK for Model Context Protocol servers and clients

A FlashAttention implementation for JAX with support for efficient document mask computation and context parallelism.