Show HN: Proxima serves 4x more requests with no hardware change on vLLM
By pythongiant · 2026-08-11 · 2 points · 1 comments
https://github.com/Tenosra/Proxima
hey everyone, i decided to make a vLLM plugin that implements the Star-KV paper. the results are quiet promising with a decode kernel thats faster than FA2 in higher batch sizes. would love any thoughts and recommendations
Open the full discussion on BetterNews