So like everybody else on the planet, I've started playing around with AI created art. In my case I downloaded Stable Diffusion and have been giving that a try. I had a few technical hurdles at first but found an easy to install version (stable_diffusion_ui) that I was able to get running on both Windows and Linux. It has a GUI which some people might prefer but the real appeal to me was that it was easy to install. Though the Linux version wasn't quit the one click install that the Windows version was, it was pretty close.
Stable Diffusion seems to work on nVidia GPUs with 4MB or more. I have seen minimum requirements listed as higher than that but as long as you stick with 512x512 images or smaller and disable Turbo it seems to work pretty well. Also, you actually don't have to have a GPU to use Stable Diffusion. It's just that if you don't, each image will take 30 minutes or more to produce depending on the speed of your CPU. It will even work with only 8 GB of system RAM but I've found it will occasionally fail if you are that limited. 16 GB is fine though. But even if that is all you have, this is open source software so you can generate as many images as you want for free as long as you are patient.
Continuing a series of steampunk themed images, this is another one I liked. I wasn't crazy about how the top of her head is chopped off here but I still like how it looks despite that. Missing whole or partial heads seems to be a common theme in AI generated art, at least with Stable Diffusion. Hands tend to be another problem though since they aren't really visible here it's not an issue.
