I've been trying to set up local SD (on windows or WSL) of and on for a few weeks now. Seems to be extremely complicated even with the packages that say they are "1 click install". Then I find out they have dependencies on Cuda, a million python machine learning packages, some package called xformers that I have to compile for my specific GPU, I have to download the right SD image and config, then I have to find and install the right upscaling package and on and on.
So no, I'm not using SD. I paid for a year of Midjourney for $100. Well worth it so far and honestly I think their latest V4 is just as good as SD. Probably is SD in fact, and it doesn't have that "Midjourney style" that previous versions had.
So no, I'm not using SD. I paid for a year of Midjourney for $100. Well worth it so far and honestly I think their latest V4 is just as good as SD. Probably is SD in fact, and it doesn't have that "Midjourney style" that previous versions had.