I was spending time in diapers last week (somehow I managed to turn a wetness indicator in a matter of minutes! Almost like I'm not potty trained at all) and it inspired me. I looked and found some rsources that helped me do LoRA training on Flux.
So I did, and my initial results are, well, amazing. There are flaws - the thing does not seem to do tapes right, and it can't nail specific diapers. But look at these:
As you can see, there's some room for improvement. But the AI does an amazing job of capturing the scene that I am going for, and I am feeling renewed confidence that I can do what I wanted to do - have a character in different diapers in different scenarios