A framework for efficient model inference with omni-modality models

[ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.

[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

TurboDiffusion: 100–200× Acceleration for Video Diffusion Models

The open source coding agent.

MotuBrain: An Advanced World Action Model for Robot Control

My learning notes for ML SYS.

Live-SWE-agent: live, runtime self-evolving software engineering agent

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

AG2 (formerly AutoGen): The Open-Source AgentOS.Join us at: https://discord.gg/sNGSwQME3x

A modular graph-based Retrieval-Augmented Generation (RAG) system

AI Agent Evaluation & Observability Platform — LangGraph + Gemini + OTEL + Grafana

🏛️ 三省六部制 · OpenClaw Multi-Agent Orchestration System — 9 specialized AI agents with real-time dashboard, model config, and full audit trails

AI agents running research on single-GPU nanochat training automatically

Official, Anthropic-managed directory of high quality Claude Code Plugins.

Model Context Protocol Servers

Let's Learn AI SYStem

Multi-Agent Harness for Production AI

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

A Low-Code MCP Framework for Building Complex and Innovative RAG Pipelines

Connect, secure, control, and observe services.

MinIO is a high-performance, S3 compatible object store, open sourced under GNU AGPLv3 license.

The Triton Inference Server provides an optimized cloud and edge inferencing solution.

The LLM Evaluation Framework

A standard API for single-agent reinforcement learning environments, with popular reference environments and related utilities (formerly Gym)

Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.

🤗 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.

verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework