News for hackers. Stuff that matters.
Re:Running a 180B parameter LLM on a single Apple M2 Ultra by passion__desire on Thursday September 07, 2023 @02:00PM (#37422981)
What if specialized smaller models is the best way ahead for community. I don't care if I am interacting with one big model which can do everything or I have to go to different websites to access specific models. All model sizes will be useful. Smaller models will be frequently used. Bigger less so.

There are no comments on this comment yet.