Hey all!
How to explain in a technical way what is Katayama_Mix_XL_v2?
Katayama_Mix_v2 is a bit like Katayama_Mix_XL_v1, but better! 😁
This new model was created by doing a 50% merge of my SDXL_Niji_Seven model, 25% of a MidJourney v6 model and 25% of a fantasy comic style model. I didn't publish these last 2 because I wasn't completely satisfied with them.
We are therefore halfway between an original model and a merge. Because the other 2 training sessions used were quite complete.
So everything is done with homemade finetuning!
This model is really very versatile. He makes very pretty images in anime, but can also do it in photos while using digital art.
-Works perfectly with euler_a and dpmpp_sde, but should have no problems with other samplers.
Have fun !
Description
FAQ
Comments (6)
From my experiments so far, it really wants to put people's hands behind their head for some reason XD But that maybe some rogue prompt word for all I know. Cool model tho, gonna share some stuff soon :3
My checkpoint is smart!
He knows that hands are not his specialty so he hides it behind the head 😌😂
I'd take "hands behind head" over "bad hands in front" any day 😂😎.
Jokes aside, I've not seen this myself. I guess it depends on the prompt.
Totally Agree it's better than previous version.
Thanks a lot 🙏
@Stan_Katayama No no i should thank you bro, Really appreciate your works.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.