Upload to ollama
#3
by
nonetrix
- opened
I am lazy
Also a Qwen2 72B based model would be awesome since it's multi lingual ability is really good! I could see myself preferring that honestly, in English it's about the same as llama3 or better as well in my personal testing. But it can speak Japanese and other languages like 29 of them only can test Japanese, but it seems at least comparable to GPT 3.5 if slightly worse by a tad but good enough to be very useful
agreed with the person above