flash-attention repositories

Trending repositories tagged flash-attention, ranked by trending score. Updated daily.

  1. 1
    xlite-dev/LeetCUDA

    An Open-source book with Modern CUDA learning Notes for Beginners - BF16/FP8/FP4, HGEMM, CuTe, Flash-Attention, etc.

    GitHub · TeX · 12,205 stars · 1,277 forks · +18 stars in 24 hours · Trending score 2.25

  2. 2
    NVIDIA/cudnn-frontend

    cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.

    GitHub · Cuda · 1,007 stars · 342 forks · +3 stars in 24 hours · Trending score 1.47

  3. 3
    ot-triton-lab/flash-sinkhorn

    The official repository of FlashSinkhorn [ICML 2026 Oral]

    GitHub · Python · 227 stars · 22 forks · +6 stars in 24 hours · Trending score 1.27

  4. 4
    joris-sta/llama-cpp-pascal

    Pre-tensor-core (Pascal GTX 1070) inference optimizations for llama.cpp: native Turbo3 compressed-KV flash attention, PTQ1 ternary DP4A matmuls, chunked gated-delta-net, TurboQuant/VBR KV on 80K-262K contexts

    GitHub · Cuda · 22 stars · 2 forks · 0 stars in 24 hours · Trending score 0.60

  5. 5
    asher/mlx-kquant

    Native K-quant support for MLX, with a quantization and fine-tuning toolchain for Apple Silicon

    GitHub · C++ · 11 stars · 5 forks · +1 stars in 24 hours · Trending score 0.56

  6. 6
    aivrar/vllm-windows-build

    Native Windows vLLM 0.33.0 (Python 3.13/CUDA 13.0), plus a Python 3.14/CUDA 13.2 build; older versions on the Releases page. PyTorch 2.14, FlashAttention/Rust, OpenAI serving, 10 KV formats, Multi-TurboQuant, experimental CPU/NVMe KV offload. No WSL or Docker.

    GitHub · Python · 68 stars · 8 forks · 0 stars in 24 hours · Trending score 0.54

  7. 7
    qflen/nsa-from-scratch

    From-scratch reimplementation of DeepSeek's Native Sparse Attention (arXiv:2502.11089) in Triton + CUDA Hopper WGMMA. 7.07x faster than FlashAttention-3 at 64k context. Five-model training fleet, perplexity sweep, LongBench v2, MoBA comparison.

    GitHub · Python · 15 stars · 0 forks · 0 stars in 24 hours · Trending score 0.36

  8. 8
    Yolande1994/CUDA_GPT2_Inference

    纯 CUDA 实现 GPT2 推理,逐步优化算子与架构,最终在单请求下推理速度领先 vLLM 约 1.5 倍。逐算子、逐迭代的进行:NCU 视察、性能解析与记录报告。

    GitHub · Cuda · 12 stars · 0 forks · 0 stars in 24 hours · Trending score 0.19

  9. 9
    IonDen/mlx-train-perf

    Fused, logit-free linear-cross-entropy loss, RAM-fit planner, and benchmark harness for MLX fine-tuning on Apple Silicon

    GitHub · Python · 6 stars · 1 forks · 0 stars in 24 hours · Trending score 0.07

  10. 10
    fishlikeX/sm70-attn

    FlashAttention brought back to Tesla V100 — a deep llama.cpp fork: SM 7.0 D256 kernels, SplitKV3, q4_0 KV cache, DFlash2 speculative decoding and multimodal fixes.

    GitHub · C++ · 19 stars · 4 forks · 0 stars in 24 hours · Trending score 0.02

  11. 11
    Indras-Mirror/llama.cpp-turboq-mtp

    Qwen4-Exp (Qwen3.8-Flash-Next) SWA + MTP + PLE-on-iswa — bounded deep-context decode w/ long-range recall, TBQ4 4.25 bpv KV, DSpark + RotorQuant. Main (master) = Qwen4-Exp build; the Qwen35 SWA hybrid lives on the qwen35 branch. RTX 4090.

    GitHub · C++ · 96 stars · 14 forks · 0 stars in 24 hours · Trending score 0.00

  12. 12
    vipshop/ffpa-attn

    Kernel Library for Large Headdim Attention (64~1024, BF16/FP8/FP4), 1.5x~15x↑ vs PyTorch SDPA.

    GitHub · Python · 341 stars · 29 forks · 0 stars in 24 hours · Trending score 0.00

  13. 13
    HKUSTDial/flash-sparse-attention

    Trainable fast and memory-efficient sparse attention

    GitHub · Python · 750 stars · 54 forks · 0 stars in 24 hours · Trending score 0.00

  14. 14
    ai-bond/flash-attention-v100

    Implementation of FlashAttention-2 for Nvidia Tesla V100 / Titan V

    GitHub · Cuda · 217 stars · 33 forks · 0 stars in 24 hours · Trending score 0.00

  15. 15
    Nathanw1014/strix-halo-llamacpp

    Performance-tuned llama.cpp for AMD Strix Halo (gfx1151): FA + MoE-prefill fixes with a bundled current Mesa driver. Vulkan and HIP; portable dir, Docker, and distrobox.

    GitHub · Python · 227 stars · 21 forks · 0 stars in 24 hours · Trending score 0.00

  16. 16
    mohamedhossammohamed/m4-prefill-engine

    Low-level Metal inference engine for Apple M4: universal quant formats at MLX parity, 1M-token out-of-core contexts, and speculative decode off flash storage.

    GitHub · Objective-C++ · 26 stars · 3 forks · 0 stars in 24 hours · Trending score 0.00

  17. 17
    alphaparkinc/genpark-flash-attention-tiling-kernel-skill

    Online Softmax Tiling FlashAttention kernel executing bounded SRAM query/key/value block updates with running max and sum rescaling for O(1) memory footprint.

    GitHub · Python · 7 stars · 0 forks · 0 stars in 24 hours · Trending score 0.00

  18. 18
    Alpha-Park/genpark-flash-attention-tiling-kernel-skill

    Online Softmax Tiling FlashAttention kernel executing bounded SRAM query/key/value block updates with running max and sum rescaling for O(1) memory footprint.

    GitHub · Python · 7 stars · 0 forks · 0 stars in 24 hours · Trending score 0.00

  19. 19
    TensorTonic/llm-optimization-handbook

    Make LLMs faster, cheaper and smaller: inference fundamentals, quantization, KV cache, speculative decoding, GPU kernels and serving, with hands-on problems.

    GitHub · 97 stars · 4 forks · Trending score 0.03