How can you do evaluation?

#2
by HoangHa - opened

Thank you for updating good models. Can I ask how I can do a benchmark for the models?

You can look at this repo for eval:

https://github.com/EleutherAI/lm-evaluation-harness

Note that there is tons of eval datasets and repos for them so you need to choose one, this was just an example :)

Wow thank you. I will test it out.

HoangHa changed discussion status to closed

Sign up or log in to comment