Svelte Hacker News logo
  • top
  • new
  • best
  • show
  • ask
  • jobs
  • about

GAAC: Lossless KV Cache Compression for Faster LLM Inference

arxiv.org

2 points by wslh 1 week ago

Kevin_Flynn 1 week ago

I searched around, but couldn't find source code. :(

  • wslh 1 week ago

    I couldn't find it either but there is something related: https://github.com/lichangshiwhu/Adacode