good work

#2
by simonko912 - opened

i usualy train normal larger models like 0.2b and they are pretty dumb, and this crossed my expectations, i usualy trained around 0.2b and a few mb to 200mb of text but i might try the opposite (smaller but more data) thanks for the idea

CompactAI org

Np, glad I could help!

Sign up or log in to comment