logoalt Hacker News

RGS1811 • yesterday at 3:34 PM • 2 replies • view on HN

Forgive the ignorance, but how are you running a 763B parameter model on 64GB of RAM?

Replies

liuliu • yesterday at 4:02 PM

SSD weights streaming. The actual weights is 552B, then engram.

➕ show 1 reply
mococa • yesterday at 3:38 PM

By “quantitizing” it

➕ show 1 reply