Expert-aware quantisation: near-Q4 quality at near-Q2 size?

3 pointsposted 11 hours ago
by martinald

No comments yet