DFloat11 [NeurIPS '25]: Lossless Compression of LLMs and DiTs for Efficient GPU Inference

CUDA Templates and Python DSLs for High-Performance Linear Algebra

This is the homepage of a new book entitled "Mathematical Foundations of Reinforcement Learning."

Lightweight coding agent that runs in your terminal

Model Context Protocol Servers

On-device AI across mobile, embedded and edge for PyTorch

The simplest, fastest repository for training/finetuning medium-sized GPTs.

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

Qwen3-Coder is the code version of Qwen3, the large language model series developed by Qwen team.

Autonomous coding agent as an SDK, IDE extension, or CLI assistant.

A high-throughput and memory-efficient inference and serving engine for LLMs

The agent engineering platform.

Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.

An elegant PyTorch deep reinforcement learning library.

A framework for efficient model inference with omni-modality models

SGLang is a high-performance serving framework for large language models and multimodal models.

The open source coding agent.

Mobile and Web client for Codex and Claude Code, with realtime voice, encryption and fully featured

Autoresearch for GPU kernels. Give it any PyTorch model, go to sleep, wake up to optimized Triton kernels.

FlashMLA: Efficient Multi-head Latent Attention Kernels

A kernel library written in tilelang

Web-based 3D visualization in Python

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

Open-source, community-driven agent harness