How a Stable Diffusion prompt changes its output for the style of 1500 artists

Post by etrautmann » Mon, Oct 03, 2022, 2:32 PM UTC

avereveard wrote: post #179

etrautmann wrote: post #58

Earlier quoted context omitted.

Yep, I see this as a start, and very curious to see the ways in which it’ll get used with a human in the loop, and also the ways human artists will be pushed to creat art that’s out of distribution for these models.

Even then, few artist got famous on technical skill alone, surely less than those who got famous primarily for their message irregardless of their skill. And this is before getting into the endless pit of defining what art is. Besides having an ai doing the legwork is no much different than Veronese giving large swath of paintings to his novices while focusing on the two/three major parts.

Post by avereveard » Mon, Oct 03, 2022, 2:59 PM UTC

etrautmann wrote: post #181

avereveard wrote: post #179

Earlier quoted context omitted.

Even then, few artist got famous on technical skill alone, surely less than those who got famous primarily for their message irregardless of their skill. And this is before getting into the endless pit of defining what art is. Besides having an ai doing the legwork is no much different than Veronese giving large swath of paintings to his novices while focusing on the two/three major parts.

I think we agree then - if new technologies allow an artist to more rapidly explore, iterate, and refine a particular message, then those artists should still have something to create beyond what is possible with these images.

Post by jhbadger » Mon, Oct 03, 2022, 4:26 PM UTC

dopidopHN wrote: post #74

jhbadger wrote: post #25

What frustrates me about Stable Diffusion is there doesn't seem to be any documentation as to what artists or vocabulary it understands. Generally people say "look at existing prompts or use various prompt generators" but that doesn't really solve the problem. I don't want to just look at what other people have randomly discovered; I want to know what the program really knows.

Honest question, would a solid understanding of the open training data help? Having the art vocabulary down as well. In effect, knowing what is present and how it’s tagged so you can « invoke » it more readily in the prompt-result. Maybe I’m out of my depth. I know the corpus of tagged image used for training is enormous … but I still think that would help the user ( a prompt-crafter )

Post by jhbadger » Mon, Oct 03, 2022, 4:50 PM UTC

bawolff wrote: post #95

jhbadger wrote: post #35

Earlier quoted context omitted.

But why isn't the list of words from those captions available anywhere (at least as far as I can tell)? There may be 600 million captions, but the number of unique words would probably be 10 or 20 thousand at most, completely feasible to browse or grep.

Its kind of a weird complaint. If i am having a conversation with someone, i wouldn't be concerned about knowing the set of all possible nouns.

Post by krapp » Mon, Oct 03, 2022, 6:09 PM UTC

otabdeveloper4 wrote: post #180

krapp wrote: post #115

Earlier quoted context omitted.

Whatever else this emergent "creative AI" phenomenon may or may not do, it's definitely touching nerves in people who still believe there's something ineffable and transcendent about the creative experience.

Stable Diffusion is literally copy-pasting existing artwork. It's not creating anything new. If anything, this sort of "AI" only makes human ineffable creative experiences even more valuable, because without it there would be no training set and no "AI".