davide-troiani/moeshift

Memory efficient and high performance Mixture-of-Experts inference for Apple Silicon

2 best points · 1 comments · 1 Hacker News post · 49 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: MoEShift - Faster large MoE inference with half the RAM (2 points, 1 comments, by davidetroiani, 2026-11-03)