Every story and comment from the Hacker News API, in a layout borrowed from 2001.
A comment on A 10 year old Xeon is all you need (for 26B-A4B MTP Drafters without GPU). Parent comment
Over 2500 tokens per second on a single request. With 8 MI300X.
There are no comments on this comment yet.
Over 2500 tokens per second on a single request. With 8 MI300X.