RIS-Kernel: Running 64k context LLMs on CPU via sparse attention

(github.com) by santosardr | May 31, 2026 | 0 comments on HN