I decided to do a big mashup of 8 of my recent SDXL models just to see how it would come out, and I got this. I wouldn't call the style super distinctive, but it's a nice anime model that renders at high detail and can do hands and bodies without any serious flaws. It's also quite good at landscapes and architecture, and follows short prompts well. I imagine it will be a good base for further training and mixing.
See this article for details about my generation settings.
Description
FAQ
Comments (6)
Damn, no one told me Christmas is on 17th of November this year.
Wow, v2 is completely different than v1. I was expecting improvement over v1 with slightly deference between same seed images but v2 generates a completely different image. I will keep both of them.
I really love this model, but somehow I get these kind of burn holes / white spots, and it seems really desaturated. Does someone know how to fix it, or where it is coming from?
Make sure you've got sdxl_vae selected and let me know if that fixes it.
Works fine now, thanks^^
I was having the same issue with this one. I know you sometimes mention when the VAE isn't baked in, but should we just assume that if you don't explicitly say that it IS baked in, that it isn't?
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.