If anyone uses anything more than the simple AI image generators, you could give this a go and might be able to start getting some decent AI gifs or better images.


https://github.com/XingangPan/DragGAN

Unfortunately it's not compatible with Stable Diffusion. GANs are an entirely different image generating architecture, they're the go-to for fake headshots but there's no general purpose pretrained models available.

I haven't played around with GANs yet. They're a viable alternative to diffusion models but running StyleGAN is beyond my meagre tech skills for now, and training a pregGAN model would probably take weeks on my hardware.

It could be useful for animating already created static images. I'll give it a go; thank you!

@gravidasomnia ImpossibleBard has the idea. I would use either but i suck and don't have the time to improve my language skills to get good stable diffusion images even. When I saw this GAN come up, i thought damn, generate stable diffusion images use them as various points but then use this the either make a rapid hyper-preg gif or to alter parts of the image that I want. As for hardware, I only have an old 980 ti, so nowhere near the memory or processing power. However, i was gonna give one of those cloud compute websites a go, but I don't feel like learning another programming skill just to make some AI art chick have juicier boobs and a more gravid belly.