Here's a model I trained for Belladonna, from the All Dogs Go to Heaven franchise.

For this model's v0.5 variant, I did something a little different. First, I manually recreated an image of Belladonna, for use on PixAI regarding something else. What I intended didn't work out, but another idea came to mind- using the image for Image-to-Image in order to clean up little issues.

The 0.5 model was trained on variations of that one picture. Anything made with that model by itself is a bit stuck in the one pose as a result. Stick with the v1.0 PixAI model for now.

As with other new models, if you want to make anything R-rated from the start, it'd be wise to save your yellow buzz and generate images elsewhere for free, then repost here.

As a reminder to the CivitAI staff, Belladonna is an outright adult character, not a minor like you mistakenly shifted her towards between February and March 2026- though I am grateful you eventually realized your error. She is shown in her various appearances as an adult, with other dogs of various ages around her proving beyond a shadow of a doubt that she is an adult.