(I'm aware it's listed as beta, but thought feedback might be useful) Seems to have good character recognition but maybe the dataset is too small as it doesn't appear to be very flexible - attempting to generate a character in a different outfit seems to almost always completely change the style or even merge two characters quite often. Going with default outfits seems to work fairly well tough some poses also seem to cause character stretching (height).

I try to use this model but it keep showing 1 model is hidden or something due to my setting i check everything tried everything but still showing one model is hidden duel to my setting and i dont know what it is

Could this be made into a Lora?

Considering that the model is based on PonyDiffusionXL V6 training, you should also use the same prompt writing method as it. It has also proven to be effective in practice.

It draws Tails significantly better than its older version. More precisely, very, very much better. I love the two-tailed fox.

If at all possible, try not to use negative prompt. In particular, negative prompt like "3D" have been shown to change the style of drawing so much that the AI generates some sort of weird style image. I'm using this conservative term because I suspect that some people might like this style.

However, if you want any "Style lora" to work, it's best not to add such negative prompt. Negative prompt like these will make the "Style Lora" almost useless.

I would like someone to make Loras to generate images, but with the art style of AOSTH, SatAM, OVA, Underground, Archie, Fleetway, Sonic X and Sonic Boom respectively.

Any chance of getting this enable to generate using the on-site feature?

whens final release?

Could this be enabled for generation on site please?

when will this be available for on site generation?

looks good i will try with some loras

Please make this available on site <3

does itt go in lora file or checkpoint ,file

I have an idea for the next version or spiritual successor of this model: In case of it's based in Pony Diffusion V6 XL, establish "source_sonic" as a trigger word instead of "Mobian or sonic character name".