News for hackers. Stuff that matters.

A comment on DeepSeek-R1. Parent comment

Re:DeepSeek-R1 by kridsdale1 on Tuesday January 21, 2025 @03:51AM (#42777847)
That would be a true Mixture of Experts.

I sometimes put the 4 biggest models like this to converge on an optimal solution

[ Parent | on HN ]

There are no comments on this comment yet.