Had a bit of a play on enhancing my previous trained stable diffusion model using LoRA. I think these are looking pretty hot now!

Let me know what you think.

This is incredible , great stuff Hope the powers in charge dont start making restrictions for Ai generated art . Like the stupid UK law that restricts my Ebike to 15.5mph .

Thank you for the kind words.

As a community we will have to make some decisions around AI generation and how the work of others is used or not used to train the models. It's probably a conversation to start sooner rather than later before someone starts magically creating new images that look just the same as the work of another producer.

I feel like the main reason this tech will do well is that the space of possible kinks far outstrips the space of viable photoshoots. Chances are people aren't going to be using this to duplicate the existing work of producers, but to extend it into their own highly specific niches. That will still be a real challenge for producer WAM, but not because of people nicking stuff. Just that people won't have to compromise on the specificity of their kinks.

Completely agree and this may well open up interesting creative options that wouldn't be possible in the real world. My point was more that the AI models will need to be trained on other images. I feel there should be some consent from producers that their images are used for this purpose rather than just being grabbed from UMD pic archive and fed into AI models.