Relative newbie here, be gentle with me.

I get really good results on Perchance, but I have a hard time replicating them when running SD locally, even when I copy/paste prompts. Perchance’s are just better , especially with Disney and Anime related subjects.

Do we know what Model/Checkpoint/LoRA’s the image generator uses, and if so, is it publicly available to download?

  • dax80@lemmy.worldEnglish
    1·
    10 months ago

    I agree. I have been trying for ages to get the same results as perchance they just seem too good. what I am not able to find is a model that is so sensitive to prompts, and can somehow do artistic outputs as well as realistic images really well. There is definitely some settings or something going on behind the scenes.