Skip to content
@flashserve

flashserve

Pinned Loading

  1. PAT PAT Public

    Prefix-Aware Attention for LLM Decoding

    Python 50 3

  2. RAGPulse RAGPulse Public

    An Open-Source RAG Workload Trace to Optimize RAG Serving Systems

    Python 41 4

  3. flash-linear-attention-npu flash-linear-attention-npu Public

    C++ 67 106

Repositories

Showing 6 of 6 repositories
  • flash-linear-attention-npu Public
    flashserve/flash-linear-attention-npu's past year of commit activity
    C++ 67 106 118 214 Updated Nov 15, 2026
  • Mosaic Public

    [ICML'26] MOSAIC: Unlocking Over 30× Context Length for Diffusion LLMs Inference via Global Memory Planning and Dynamic Peak Taming

    flashserve/Mosaic's past year of commit activity
    Python 10 Apache-2.0 0 0 0 Updated Nov 8, 2026
  • ReWY Public

    Faster Gradient Computation for Delta-Rule Attention Family

    flashserve/ReWY's past year of commit activity
    Python 11 MIT 0 0 0 Updated Nov 3, 2026
  • OLED-MoE Public

    Accelerating MoE-Based dLLM Inference via Inter-Iteration Locality-Aware Expert Offloading

    flashserve/OLED-MoE's past year of commit activity
    Python 9 0 0 0 Updated Sep 30, 2026
  • RAGPulse Public

    An Open-Source RAG Workload Trace to Optimize RAG Serving Systems

    flashserve/RAGPulse's past year of commit activity
    Python 41 MIT 4 0 0 Updated Sep 1, 2026
  • PAT Public

    Prefix-Aware Attention for LLM Decoding

    flashserve/PAT's past year of commit activity
    Python 50 MIT 3 0 0 Updated May 26, 2026

Top languages

Python C++