Re:How a Stable Diffusion prompt changes its output for the ...
by daveguy on Sunday October 02, 2022 @05:02PM
(#33060234)
The positioning seems very consistent, almost to the point where I wonder if that was part of the selection process to demonstrate the differences in style. There are only four per style, where the position of a subject could be a selection factor. Hard to tell if the position similarities are driven by the Stable Diffusion model or by the selection of representative images.
Re:How a Stable Diffusion prompt changes its output for the ...
by tough on Sunday October 02, 2022 @05:18PM
(#33060383)
The composition and positioning come from the original seed. If the same seed is used, the same background image noise is applied for the initial image which is transformed into all the styles.
Thus the similarities you see would make sense if using also the same seed for the tests.
Re:How a Stable Diffusion prompt changes its output for the ...
by totetsu on Monday October 03, 2022 @03:43AM
(#33064588)
Yes, if you generate a bunch of images of waves or the sea, or other repeating patterns with the same seed, you can see how all the 'peeks and toughs' of those patterns line up in the same place.
Re:How a Stable Diffusion prompt changes its output for the ...
by olejorgenb on Sunday October 02, 2022 @05:16PM
(#33060364)
Maybe it's based on img2img? Some are different enough that it's not obvious that's the way it has been done though.
Re:How a Stable Diffusion prompt changes its output for the ...
by adityashankar on Sunday October 02, 2022 @05:58PM
(#33060717)
Hey, I'm the developer of this project, no this is not based on img2img all the images just have the same seed
Re:How a Stable Diffusion prompt changes its output for the ...
by michael-ax on Sunday October 02, 2022 @06:06PM
(#33060794)
incredible job! i was just showing my own experiments with sd to a friend and this just take the cake. thank you so much for mixing in that artist list!!
Re:How a Stable Diffusion prompt changes its output for the ...
by nl on Sunday October 02, 2022 @08:00PM
(#33061729)
If you play with Stable Diffusion enough this behavior becomes very apparent. Changing the seeds will give different results, but even relatively significant changes in the prompt will still find similar themes or layouts.
Re:How a Stable Diffusion prompt changes its output for the ...
by djmips on Sunday October 02, 2022 @03:59PM
(#33059596)
So interesting, and really seems somehow related to how we dream, hallucinate.. or even experience reality?
Re:How a Stable Diffusion prompt changes its output for the ...
by deltasevennine on Sunday October 02, 2022 @04:47PM
(#33060090)
Arguably it could be different from our experience. It could even be a superior and more efficient methodology then the things our brain uses to imagine things.
Re:How a Stable Diffusion prompt changes its output for the ...
by kache_ on Sunday October 02, 2022 @05:16PM
(#33060366)
from "is your blue my blue" to "is the AI's blue our blue" :)
Re:How a Stable Diffusion prompt changes its output for the ...
by xapata on Sunday October 02, 2022 @09:36PM
(#33062433)
The blue woman in Sam Bosma's style looks strikingly similar to the one in Noah Bradley's style. They even have the same 3 patches of pink on their cheeks.
https://static.adityashankar.xyz/gorgeous/hair_flowers_full/...
https://static.adityashankar.xyz/gorgeous/hair_flowers_full/...
For example, I flipped back and forth between Beatrix Potter and Paulus Potter. A rounded white bonnet in one picture becomes a couple of blossoms in the other. The roof of a house becomes some shadowy wall with plants in the other. Two flower pots are very similar, just with slightly different coloring.
It makes it more apparent that the algorithm etches the images out of noise, and if the seed is the same for two images with different prompts, you're likely to see traces of that noise represented differently but recognizable in both images.