Appreciate all the effort put into this! It seems like it might benefit from some refining, as it appears a bit over baked. I've been grappling with regional prompting and inpainting and think I'm happy with the result.

I've noticed that the background consistently appears blurry, which suggests that much of your training data may have depth of field. This makes it challenging to prompt specific background elements. (I had to disable your Lora to Inpaint a background.) Additionally, it tends to override any original art style present.

Want to say thank you for all the work though, I really like that some of the generations include a spitting motion which I think the saliva trails prompts alone can struggle with.

Your images are impressive! Keep up the good work.