DeepSeek Harness: Everything is a Plugin.

A Minimal and Elegant Framework & Tutorial for Real-Time Interactive World Models

Language model tokenization at GB/s

A hive mind communication platform

Build your hardware, easily!

MobileFineTuner: Native C++ framework for fine-tuning LLMs directly on mobile devices. Features: LoRA/Full-FT, ZeRO-inspired parameter sharding, energy-aware throttling, custom autograd engine. Kee…

Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.

Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦

Publish Vega visualizations as embedded web components with interactive parameters.

Tile-Based Runtime for Ultra-Low-Latency LLM Inference

The Open Framework for Building Agents

A tool for creating and running Linux containers using lightweight virtual machines on a Mac. It is written in Swift, and optimized for Apple silicon.

Efficient Triton Kernels for LLM Training

The codebase of Cola DLM

A Python DSL to write Nvidia PTX for Hopper and Blackwell in JAX and PyTorch

repo for Numba-CUDA-MLIR

A curated list of papers, tools, and resources on Multi-Token Prediction (MTP) and related techniques in Large Language Models (LLMs), Speech-Language Models (SLMs), and more.

Typescript/React Library for AI Chat💬🚀

The agent that grows with you

xURL: Client for AI Agents URLs

The Vampire Theorem Prover

Remove unused Rust dependencies with this one weird trick!

Launch apps by casting spells! 🪄

Linux virtual machines, with a focus on running containers

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends

Python PyPi staging server and packaging, testing, release tool

⚡ A fast Git hook manager written in Rust, designed as a drop-in alternative to pre-commit, reimagined.

JavaScript library for conversion between TeX/LaTeX and Typst math code.

An open-source AI agent that brings the power of Gemini directly into your terminal.