creation123Q: I have a few questions: 1) What version of Stable Diffusion do you use?
If you use 2, will it be possible to also use older versions in the future?
2) Which Stable Diffusion models have your incorporated (e.g., F222, Hassanblend, pixel art - SD2, etc...)?
3) Will you add the ability to generate 4 images at a time?
Rexxar
May 14, 2024A: Hey there,
Thanks for the questions. We're using 2.1 as the vanilla model. We are bringing the best custom models from 1.5 and elsewhere into the platform and will continue to do this. We hope to also offer the raw 1.5 alongside the 2.1 in the future but see the future of the technology in the customization.
Currently have the Midjourney V4, Redshift, Analog, Anything V3, Tron, Robot, Mo Di and a few others. Those you mentioned are possible to add, we just have to be cautious if we do so. We're not against users creating NSFW images with Supermachine but we take a zero tolerance to anything that could be perceived as illegal so it is quite a process to add NSFW images and safeguard in that way to protect ourselves, we're working on this area.
Multiple generations is something we look at, in Supermachine, you don't have to wait for an image to finish generating before you hit generate again so whilst it isn't hitting a number and hitting generate, it is fast to queue up like that. We're exploring other ways of course.