Flash Prefill

Accelerates long-context LLM prefilling by identifying sparse attention patterns without expensive scoring, using block-level approximations and dynamic thresholding. Achieves 27.78x speedup at 256K tokens while maintaining accuracy.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit e8847dd492

Frequently asked questions

npx get-skill add adu2021/flash-prefill