I use outdated technology just for fun, listen to crappy music, and watch a lot of horror movies. Expect posts about These Things. I talk a lot.
Check tags like Star Trek Archive and Media Piracy to find things I share for others.
i forked up for a month of deluxe sub that lets me do infinitely many lower-res generations, and i'm gonna get my money worth out of it (in terms of shitposts)
is the training set available for download? i've been running stable locally and that would be the Ideal
i know the NAI anime model has been leaked! but given that i'm on windows with a radeon rx 580 (thanks shitcoiners), playing with it locally is very impractical, so i never tried it
Letting people here know of the Stable Horde service which is basically free* AI image generation. I suggest Stable UI https://aqualxx.github.io/stable-ui/
(* It's all based on a karma system. If you have the GPU power, you can make a "worker" that contributes to the horde. It's all crowdsourced!)
Really cool! It's obviously way slower but it gives a ton more tools. If you're familiar with the set of more limited controls NAI Diffusion exposes, any interesting high-impact tools to play with?
Just wondering what specific options are particularly interesting to experiment with coming from NAID. Given its response speed I can't simply try them all to figure it out lol
If you use the anonymous account, you basically put yourself way in the back of the line (Which, by the default, that's what it uses), so it's gonna take a bit to goto the front of the queue during high load.
That said, I think playing around with guidance might be fun to try.
Theory: Hatsune Miku and Uwawawawawwaawaawaawawwawawa don't appear together for the same reason superman and clark kent don't
I tend to feel an aversion to anything to do with AI art, but this was very interesting. It's sort of like an SCP.
if you date a gal with Uwawawawawwaawaawaawawwawawa as a symbiote, are you technically polyamorous