DeepSeek-v4.1 Flash: Pushing the Limits of KV Cache Compression

(zartbot.github.io)

73 points | by mfiguiere 3 hours ago ago

6 comments

  • arikrahman 2 hours ago

    I am very impressed with the KV Cache Compression work as well as the prefix cacheing making queries converge on practically free.

  • vivzkestrel an hour ago

    404 on the blog page? https://zartbot.github.io/blog/

  • smy20011 an hour ago

    Removed

    • girvo an hour ago

      The website design definitely is, but I don’t know if the content is? This reads pretty human to me and is quite interesting to boot!

    • yunfei an hour ago

      you don't know him?

    • sebmellen an hour ago

      The writing feels human to me… and I call out AI slop as much as possible.