About this version
512*768 !!!
I recommend 0.7
Trigger words : holding phone,cellphone picture,nudeP,1girl,solo,
AOM3
Not perfect success, only 50% or 25%, you need try more times.
and you need ADetailer to fix face.
if you have more patience , you can try more things.
I just know it can work, Maybe it can do better.
True trigger words : holding phone,cellphone picture,nudeP
If you find that there are no hands or phones on the picture, you can increase weight.
If you find multiple girls in the picture, you can decrease weight.
it's 0.1 version .I know it's not good.
it's more complicate than I thought.
Description
FAQ
Comments (4)
Great stuff! Awesome ❤️
Would be awesome to get a trigger word for live viewport with recording icon on that accurately reproduces the general subject in the picture from slightly different angle to match the phone position....
It would be awesome but it's also one of the hardest things you could ever ask Stable Diffusion to do. SD has no idea what it means for things to exist in physical space and doesn't really understand what "subjects" are. It only knows what kinds of visual patterns tend to show up when certain words are used.
I imagine someday we'll get an image generation model that actually can "visualize" things, but it won't be SD. It's gonna be a while. 🙁
definitely need one for pony bro, this model could do so much in it!