Meathead 1173 it is interesting No no i

Yeah just read this on a GIt.

I'm running LLaMA-65B on a single A100 80GB with 8bit quantization. $1.5/hr on vast.ai

The output is at least as good as davinci.

@Meathead 30B works on our bot now

Yeah I just got it working myself 😄

You dont any tests with it yet?

done"

What I noticed is that it acts more like Dav

Also people are saying it runs faster on linux then windows.

It’s running 1it/s on the 16bit version @Meathead