• Alone or with a partner, explore the creative application of img2img techniques in Stable Diffusion together with p5 based on these starter codes. (Note: the result of this doesn’t need to be interactive - it could e.g. also be a static animated GIF created using code!)

<aside> <img src="/icons/chat_lightgray.svg" alt="/icons/chat_lightgray.svg" width="40px" /> sketch here!!! ☜

generate GIF based on only one black dot! The major change I made is to generate pictures based on the one we just received from SD. In that case, the graphic is keeping changing and I assume we can see more variations.

（Oh no the GIF is too big, if it don’t work please try this link)

https://drive.google.com/file/d/17KuGOEofk9lnW7njfLFvbOoaLxyDdEp6/view?usp=drive_link

</aside>

Some Questions

▶︎ For week 9’s sketch, because I keep generating based on what SD send to me, I find they all end in similar way, with those heavily vague outline, fluorescent green，and weird blue that diffuses outside the outline. Maybe it has something to do with the way they generate new graphics based on the old one. And also the way they deal with colors and training data.

▶︎ I meet the same problem with week 8’s sketch. If I ask SD to generate a picture for me, sometimes the subject will not be fully shown on the canvas. I’m wondering if it is the problem with the way I place my canvas and SD’s pictures, or it is the problem with the SD’s pictures (like they originally look in that way)

(No shoes shown even though I asked for “fully body”)