Quantisation script

#1
by Rijgersberg - opened

Hi Robin, very nice work! Having fun playing around with the model right now.

One quick question: do you have a script that produces all these quantisation levels with the "correct" settings? Or is it just manual llama.cpp work?

Thanks! πŸ˜€

Yes, i have a script for that. I will finish it tonight and add it to my GitHub repo.

Will add a comment once it is added.

Conversion script has been added to the github repo: https://github.com/RobinSmits/Dutch-LLMs

It is very basic...but it works...

robinsmits changed discussion status to closed

Sign up or log in to comment