scrya-com/rotorquant

KV cache compression via block-diagonal rotation. Beats TurboQuant: better PPL (6.91 vs 7.07), 28% faster decode, 5.3x faster prefill, 44x fewer params. Drop-in llama.cpp integration.

2 best points · 1 comments · 1 Hacker News post · 1,057 GitHub stars · View on GitHub

Hacker News discussions

• RotorQuant: Faster Than TurboQuant (2 points, 1 comments, by Klaster_1, 2026-04-19)