Here is a video created by animating the scale parameter in Stable Diffusion:
This uses the prompt "a photo of San Francisco shrouded by fog 8k landscape trending on artstation", seed 56345, dimensions 896x448 and frames that range from scale 0.0 to 21.9. The higher the scale the more literally your prompt should be interpreted.
Interesting at scale 0 Stable Diffusion seems to think San Francisco is a diverse crowd of people protesting something (and it's not wrong):
From there is settles on a foreground quickly, experiments with various numbers of Transamerica Pyramids and then starts throwing in various permutations of the Golden Gate Bridge. It eventually decides that the Transamerica Pyramid belongs in Marin and that the bridge is so iconic that the city is probably filled with red aircraft warnings lights.
I think the versions a third of the way in are the best.
- A black lab chases a Roomba and then things start to get weird...
- Stable Diffusion Watches a Security Camera - a short horror movie
- Style Transfer for Time Lapse Photography
You Might Also Like
I'm using the command line version of stable diffusion. To generate the images I wrote a script that just output a batch file of the command line for each scale step. I used DDIM, ddim_eta was 0 though.
This is interesting, thanks for posting it. May I ask, did you have a script or something to automate/batch the parameter? or did you make all the frames manually? Also, I'm curious what sampler you used, because if I'm correct, DDIM with eta adds some extra randomness to the final output. I just learned this after making note of a fig ton of my favorite seeds, which I now find are useless because I'd used a high eta level.