rip, just from looking at the parameter size of the new OSS models from OpenAI, I don’t think I’ll be able to get either to run on my hardware 😭

(They have a 20b model & a 120b one)