undefined | Better HN

0 pointsilaksh11h ago0 comments

Could be amazing, but it's hard to judge if it will really work with say a 27 B model or larger. We can already get pretty good speed with a 2B model.

0 comments

gaeld11h ago

thanks! we explain how it scales to larger models in the last section the OP blog post

bcjdjsndon7h ago

Shame you stopped short of actually benchmarking that scale though, eh?

gaeld6h ago

will do - we are a small team and it takes time to implement and optimize a new model, whatever the size.

2 more replies

j / k navigate · click thread line to collapse