It's probably my second most used model so I figured it was worth updating with some slight changes now that I understand a little better how it works.

@Civitai_Rom Oh you don't have to worry about that, I think it's something all of my models will do even if I don't tag for them during training since the images in the datasets have one or three hearts in them. They'll probably continue sneaking in when you least expect it.

The training data was in a soft watercolor style that Illustrious seemed to have some trouble absorbing properly, so it took a bit of testing different epochs to find the right balance to keep that look without forcing foxes too much into images, but I think this one I uploaded looks about right. Does seem to have a bit of trouble making rabbit ears, though.

Thanks a lot for giving it try, it's a really simple model but it likes to nudge generations in a way that I couldn't get any checkpoint to do with just regular prompting, but it could be that I just didn't know how to do it properly.