My first attempt at doing an SDXL/Pony model merge, this one is a mixture of Hassaku XL, Honey Mix [XL] High contrast, and a tiny injection of Indigo Furry Mix XL. Very versatile, but has a heavy emphasis on anime styling (I struggled to get realistic or 3D generations myself, so I'll call it an inherent limitation.) It's able to do a wide variety of characters, settings, poses, costumes, and digital art styles, including character sheets. It can also tackle anthropomorphic characters rather well.
As this merge contains Hassaku XL, it inherits the permissions related to that model, as well as its license information.
License info:
Because this model is a continuation of pony, from PurpleSmartAI, bears the following license:
This model is licensed under a modified Fair AI Public License 1.0-SD (https://freedevproject.org/faipl-1.0-sd/) license.
The following modifications have been added to Fair AI Public License:
You are not permitted to run inference of this model on websites or applications allowing any form of monetization (paid inference, faster tiers, etc.). This applies to any derivative models or model merges.
Description
Updated version of merge, many improved generations removed from its forefather. Higher detail, more versatility, greater anatomical accuracy. This one tends a bit more toward digital art/anime styled imagery, while retaining the improvements of the previous version, and has a bit more expressive facial detail.
FAQ
Comments (3)
Very positively surprised with how well it does furry males, unlike many other models. Great prompt understanding, even with some more specific descriptions, locations make even more sense then in other pony models i feel like. Already noticed huge improvements in the new version compared to the first of this too, it now feels like it pretty much has the entire realism-model slapped into it too and can now do realism and art-styles very nicely, and even mix them.
Only small gripes i have, this and your other pony merges dont know even remotely what a Greninja is. c': And that (furry) artist-tags dont seem to get recognized at all, im guessing theyre just not included, i think ive read from other models specifically taking them out, idk what the case is here.
If you use the Forge UI, you can get even better results with the new DDPM and HeunPP2 samplers; I've been testing those out for a day or two and been very satisfied with the results so far.
I tend to revert to some of my other older models as part of a workflow to capture specific artstyles, since XL and Pony-derived models struggle with it, deliberately it would appear, without LORAs or embeddings to handle it.
@AI_nstein Interesting, i will try those, ive just stuck to DPM++ 2M SDE and then now 3M SDE for now. I always forget to try out the others, once i start generating i just dont remember thats an option i can change too.