It seems to be a Chat finetune
#1
by
adamo1139
- opened
Hi,
I was hoping this was a llama-fied base Qwen1.5-32B but it does respond to chatml prompting in a way that I would have expected from Qwen/Qwen1.5-32B-Chat. Can you please clarify if it's based on Qwen/Qwen1.5-32B or Qwen/Qwen1.5-32B-Chat?
hi this is base model
Thanks for confirming. I guess Qwen team trained their model on a lot of ai assistant data before releasing it as "base" then.
adamo1139
changed discussion status to
closed