GStars
    fla-org

    fla-org/flash-linear-attention

    🚀 Efficient implementations for emerging model architectures

    llm
    nlp
    large-language-models
    machine-learning-systems
    natural-language-processing
    sequence-modeling
    Python
    MIT
    6.0K stars
    786 forks
    6.0K GitHub watchers
    Updated Oct 26, 2026
    View on GitHub

    Build with Backblaze B2

    SDKs, agent skills, IDE extensions, and reference pipelines from Backblaze Labs. All open source.

    Explore Backblaze Labs

    Loading star history...

    Where it stands

    Rank, growth pace and release cadence relative to everything GStars tracks.

    1 release in the last 90 days, the latest 3 days ago, with 168 contributors on GitHub.

    Gained 11% more stars in the last 90 days than in the 90 days before.

    Use Cases & Benefits

    • Provides efficient Triton-based PyTorch implementations of state-of-the-art linear attention models for large-scale sequence modeling.
    • Delivers platform-agnostic, memory-efficient kernels compatible with NVIDIA, AMD, and Intel hardware, enabling faster training and inference.
    • Use for integrating advanced linear attention mechanisms into custom large language model architectures to improve scalability and speed.
    • Use for benchmarking and comparing emerging linear attention models in research on natural language processing and sequence modeling.
    • Use for developing hybrid or fused modules that optimize memory usage and computational efficiency in transformer-based training pipelines.

    Should you depend on it?

    License terms, project hygiene and risks, from GitHub metadata and GStars' snapshots.

    License

    MIT

    Permissive: commercial use, modification and redistribution are allowed with attribution.

    Community files

    • README: present
    • Contributing guide: present

    Watch-outs

    No watch-outs from GStars' checks.

    General guidance from repository metadata, not legal advice. How we score

    From the README

    This repo aims at providing a collection of efficient Triton-based implementations for state-of-the-art linear attention models. All implementations are written purely in PyTorch and Triton, making them platform-agnostic. Currently verified platforms include NVIDIA, AMD, and Intel. Any pull requests are welcome!

    Check out the PR and their blog post for more infos!

    Roughly sorted according to the timeline supported in fla. The recommended training mode is chunk when available.

    Excerpt from the fla-org/flash-linear-attention README on GitHub (MIT license). © its authors.Read the full README on GitHub

    Daily snapshots

    Stars gained per day, and how forks and open issues moved, from GStars' daily snapshots

    Jul 30: +14Jul 31: +3Aug 1: +0Aug 29: +8

    +511 stars over 90 days across 11 snapshot days; best day +14 on Jul 30.

    Forks

    786+169

    Open issues

    133+54

    Compared with peers in LLM

    Fastest-growing repositories GStars tracks in this cohort for the week of Nov 1, by 7-day star growth

    View category →
    fla-org/flash-linear-attention compared with LLM repositories by stars, 7-day star growth, health score and last release
    RepositoryStars+7dHealthLast release
    thedotmack/claude-mem113.6K+8.8K35today
    NVIDIA/auto-ontology7.5K+5.6K37—
    HKUDS/DeepTutor55.5K+5.5K655 days ago
    affaan-m/ECC297.9K+5.2K6335 days ago
    DietrichGebert/ponytail185.3K+5.1K607 days ago
    fla-org/flash-linear-attention6.0K—603 days ago
    Category median113.6K+5.5K60—
    Compare flash-linear-attention with claude-mem, auto-ontology, DeepTutor →

    Discover Repositories

    Search across tracked repositories by name or description