Repositories list

• flash-linear-attention

  Public
  🚀 Efficient implementations of state-of-the-art linear attention models

  natural-language-processingmachine-learning-systemslarge-language-models

  Python

  •

  MIT License

  •381•4.4k•47•17•Updated Feb 9, 2026Feb 9, 2026

• native-sparse-attention

  Public
  🐳 Efficient Triton implementations for "Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention"

  Python

  •

  MIT License

  •49•964•4•1•Updated Feb 5, 2026Feb 5, 2026

• flash-bidirectional-linear-attention

  Public

  fla-org/flash-bidirectional-linear-attention’s past year of commit activity
  Triton implement of bi-directional (non-causal) linear attention

  computer-visionmachine-learning-systemstriton-lang

  Python

  •

  MIT License

  •2•65•1•0•Updated Feb 2, 2026Feb 2, 2026

• hybrid-distillation

  Public
  Python

  •4•25•2•0•Updated Dec 31, 2025Dec 31, 2025

• flame

  Public
  🔥 A minimal training framework for scaling FLA models

  sequence-modelinglarge-language-modelsflash-linear-attention

  Python

  •

  MIT License

  •54•344•2•5•Updated Nov 15, 2025Nov 15, 2025

• fla-zoo

  Public

  fla-org/fla-zoo’s past year of commit activity
  Flash-Linear-Attention models beyond language

  computer-visionflash-linear-attention

  Python

  •

  MIT License

  •0•21•0•0•Updated Aug 28, 2025Aug 28, 2025

• distillation-fla

  Public
  Distillation pipeline from pretrained Transformers to customized FLA models

  Python

  •

  Apache License 2.0

  •7•11•0•0•Updated Aug 4, 2025Aug 4, 2025

• fla-eval

  Public
  A minimal evaluation framework for FLA models

  0•7•0•0•Updated Aug 3, 2025Aug 3, 2025

• vllm

  Public

  fla-org/vllm’s past year of commit activity
  A high-throughput and memory-efficient inference and serving engine for LLMs

  Python

  •

  Apache License 2.0

  •13k•1•0•0•Updated Jul 7, 2025Jul 7, 2025

• fla-synthetic-kit

  Public
  Painless Evaluation of Flash Linear Attention models on Synthetic Tasks

  0•5•0•0•Updated May 13, 2025May 13, 2025

• fla-rl

  Public
  A minimal RL frame work for scaling FLA models on long-horizon reasoning and agentic scenarios.

  MIT License

  •0•5•0•0•Updated Apr 1, 2025Apr 1, 2025

• ThunderKittens

  Public
  Tile primitives for speedy kernels

  Cuda

  •

  MIT License

  •236•2•0•0•Updated Mar 27, 2025Mar 27, 2025

• flash-hybrid-attention

  Public
  0•7•0•0•Updated Mar 5, 2025Mar 5, 2025