Throwaway4669332255

But i'll trickle down? Right? RIGHT??

Idk man I've yet to know anyone who died from drinking magma.

How does the Nemo 12B compare to the Llama 3.1 8B?

Apparently I am an idiot and read the wrong paper. The previous paper mentioned that "comparable with the 8-bit models"

https://huggingface.co/papers/2310.11453

They said their's is "comparable with the 8-bit models". Its all tradeoffs. It isn't clear to me where you allocate your compute/memory budget. I've noticed that full 7b 16 bit models often produce better results for me than some much larger quantied models. It will be interesting to find the sweet spot.

So are more bits less important than more paramters? Would a higher paramter or higher bit count matter more if the models ended up the same size?

I'm so glad I work for a medium-small company. We moved to a smaller office and only require to go in twice a month

Thank you! I had no idea this existed.

I am surprised reddit hasn't removed this post yet.

I got an account banned for saying "lemmy dot world" when someone asked "Are there even any good alternatives?"

He's opinionated but a pretty good science communicator.