Show HN: Proxima serves 4x more requests with no hardware change on vLLM https://ift.tt/NolYc8x

Show HN: Proxima serves 4x more requests with no hardware change on vLLM 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 https://ift.tt/IKDHYO7 August 11, 2026 at 10:04PM

Komentar

Postingan populer dari blog ini

Launch HN: PillarPlus (YC W20) – Automatically create construction blueprints https://ift.tt/2yet5m3

Support San Francisco Small Businesses