Skip to content
@flagos-ai

FlagOS

A Unified, Open-Source AI System Software Stack
README.md

FlagOS: A Unified, Open-Source AI System Software Stack

GitHub License

Language: English | 中文

📖 About FlagOS

FlagOS is a unified, open-source AI system software stack designed for multi-chip scenarios. It was jointly initiated and established by over ten domestic and international organizations, including chip companies, system manufacturers, algorithm and software-related entities, non-profit organizations, and research institutions.

Addressing core pain points in utilizing diverse AI chips, FlagOS builds a comprehensive system software ecosystem that demonstrates the potential to break down ecosystem barriers between different chip software stacks, effectively reducing migration costs for developers.

🎯 Core Components

Component Purpose Repository
FlagGems High-performance universal AI operator library FlagGems
FlagTree Unified AI compiler FlagTree
FlagScale Unified parallel training and inference framework FlagScale
FlagCX Unified communication library FlagCX
FlagPerf Multi-chip evaluation tool FlagPerf
FlagRelease Large model release platform FlagRelease
Triton-Copilot AI enabled operator develop tool Triton-Copilot

🚀 Extended Ecosystem Components

Developer Tools & Utilities

  • TritonCopilot: Next-generation AI-assisted kernel engineering for multi-chip systems → TritonCopilot

  • libtriton_jit: Triton JIT C++ runtime for reduced Python overhead → libtriton_jit

Framework Enhancements

  • Megatron-LM-FL: GPU-optimized library for training transformer models at scale → Megatron-LM-FL

  • TransformerEngine-FL: FP8 mixed precision training for transformer models → TransformerEngine-FL

Inference & Serving

  • vllm-FL: A vLLM plugin built on FlagOS unified multi-chip backend for LLM serving → vllm-FL

  • vllm-plugin-FL: Advanced vLLM plugin for enhanced multi-chip support → vllm-plugin-FL

Community & Resources

  • EasyOfUse: Simplify development for open compute - plug-and-play solutions and best practices → EasyOfUse

  • Community: FlagOS community governance, contribution guides, and communication hub → Community

📚 Documentation & Resources

Knowledge Base

Model Repositories

Community Channels

  • WeChat Official Account: FlagOpen
  • WeChat Channels Account: FlagOpen

🌐 Join the Community

Ready to contribute or learn more? Visit the FlagOS Community Repository for:

  • 📖 Comprehensive contribution guidelines
  • 💬 Community discussions and support
  • 🤝 Ways to get involved
  • 📚 Resources and documentation

💕 Contributing

We welcome contributions from the community! Whether through starring, forking, or submitting pull requests, your involvement helps make FlagOS better.

  • ⭐ Star - Show your support!
  • 🔔 Watch - Stay updated with our latest releases
  • 🍴 Fork - Start contributing!

Code of Conduct

Please read and respect our Code of Conduct:

📄 License

FlagOS and its components are licensed under the Apache License 2.0. See individual repositories for details.

🤝 Partners & Contributors

FlagOS is jointly supported by leading organizations in chip manufacturing, system software, and AI research. This collaborative effort demonstrates the commitment to building an open, unified ecosystem for AI system software across diverse hardware platforms.

Join us in building the future of AI system software! 🚀

Pinned Loading

  1. FlagGems FlagGems Public

    FlagGems is an operator library for large language models implemented in the Triton Language.

    Python 809 182

  2. FlagScale FlagScale Public

    FlagScale is a large model toolkit based on open-sourced projects.

    Python 430 130

  3. FlagAttention FlagAttention Public

    A collection of memory efficient attention operators implemented in the Triton language.

    Python 287 19

  4. flagtree flagtree Public

    FlagTree is a unified compiler supporting multiple AI chip backends for custom Deep Learning operations, which is forked from triton-lang/triton.

    C++ 148 31

  5. community community Public

    FlagOS community governance, developer contribution guide, contribution agreement, and community communication | 包含FlagOS社区治理、开发者贡献指南、开发者贡献协议、社区交流等内容

    4 1

  6. EasyOfUse EasyOfUse Public

    Repo to collect easy of use issues for all FlagOS components | 收集FlagOS易用性问题，欢迎多提建议。

Repositories

Showing 10 of 18 repositories
  • FlagCX Public
    flagos-ai/FlagCX’s past year of commit activity
    C++ 137 44 2 9 Updated Dec 25, 2025
  • flagtree Public

    FlagTree is a unified compiler supporting multiple AI chip backends for custom Deep Learning operations, which is forked from triton-lang/triton.

    flagos-ai/flagtree’s past year of commit activity
    C++ 148 MIT 31 7 (3 issues need help) 11 Updated Dec 25, 2025
  • FlagGems Public

    FlagGems is an operator library for large language models implemented in the Triton Language.

    flagos-ai/FlagGems’s past year of commit activity
    Python 809 Apache-2.0 182 65 (2 issues need help) 75 Updated Dec 24, 2025
  • FlagOS-Robo Public

    FlagOS-Robot provide end-to-end toolkit for embodied data loading, model training, inference, and evaluation boosting the iteration efficiency of the data-model loop and accelerating both the development and industrial adoption of embodied intelligence technologies.

    flagos-ai/FlagOS-Robo’s past year of commit activity
    0 Apache-2.0 0 0 0 Updated Dec 24, 2025
  • vllm-plugin-FL Public
    flagos-ai/vllm-plugin-FL’s past year of commit activity
    Python 3 3 0 1 Updated Dec 24, 2025
  • FlagScale Public

    FlagScale is a large model toolkit based on open-sourced projects.

    flagos-ai/FlagScale’s past year of commit activity
    Python 430 Apache-2.0 130 7 59 Updated Dec 24, 2025
  • libtriton_jit Public

    A Triton JIT runtime and ffi provider in C++

    flagos-ai/libtriton_jit’s past year of commit activity
    C++ 29 MIT 10 2 0 Updated Dec 23, 2025
  • FlagRelease Public
    flagos-ai/FlagRelease’s past year of commit activity
    1 Apache-2.0 1 0 0 Updated Dec 19, 2025
  • TransformerEngine-FL Public Forked from NVIDIA/TransformerEngine

    A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.

    flagos-ai/TransformerEngine-FL’s past year of commit activity
    Python 3 Apache-2.0 592 0 1 Updated Dec 17, 2025
  • Megatron-LM-FL Public Forked from NVIDIA/Megatron-LM

    Ongoing research training transformer models at scale

    flagos-ai/Megatron-LM-FL’s past year of commit activity
    Python 0 3,480 0 0 Updated Dec 15, 2025
View all repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Python C++

Most used topics

Loading…