Text Generation
Transformers
Safetensors
llama
Not-For-All-Audiences
nsfw
text-generation-inference
Inference Endpoints
Settings
#2
by
CulturedMan
- opened
When I used the settings I had for Utopia on this one it outputted nonsense, but upon changing those settings it proved to be a great model. It seems UtopiaXL operates best at a lower temperature alongside tfs/top_p while Utopia does better with more freedom.
That's my experience anyway.
have you used min p at all?
I'm using Sillytavern, and I don't see a 'min p' option there. There is 'typical p' though. I haven't touched that yet.