Anyone have any good tricks for getting AI image generation models like #dalle or #stablediffusion to produce animals or people with three eyes? I was hoping to get a lemur with a typical "mind's eye" third eye, but all the models seem to ignore the third eye condition no matter how frequently I specify it.
Tried to go extremely stereotypical, specifying the "three eyes" multiple ways. The closest I got was a lemur with a little yellow light on its forehead.
“a three-eyed lemur; the lemur is sitting in a lotus position and all three of its eyes are open. The third eye is centered on its forehead. highly detailed digital art trending on artstation”
@pganssle Maybe mention three-eyes terms: trepannation, minds eye, Ajna chakra, …. The same terms any consumed art would have used in their description.
@adamchainz Yeah, tried mind's eye, lots of terms like that. Just tried Ajna chakra and I got a bunch of lemurs with glowing balls at all their *other* chakras...
@pganssle Damn… img2img might work. Badly photoshop what you want and iterate from there. I’ve achieved some okay results with that feature on DiffusionBee.
Btw, why? Some weird D&D campaign or something?
@adamchainz I am putting together an Anki deck to teach my son common sight words (to supplement our phonics-based reading). I have a sight word for him to read, then a simple example sentence, and when we finish the sentence it reveals an illustration.
I had the sentence "I have three eyes" as an illustration for "have", and I wanted to have a three-eyed lemur or something (kids love funny stuff like that), but no matter what I tried I couldn't get any image generator to give me something suitable.
I ended up going with a different example sentence, but I found it curious that image generators *really* don't like extra eyes.
@adamchainz I think if I weren't concerned with licenses I could just google image something up pretty easily, but I'm trying to keep all the images CC-0 or CC-BY so that when the deck is done I can open source it. I am considering trying to go on fiverr or something to outsource the illustrations, because they're really quite time consuming even using generative AI, but unless I can find someone who will churn out a ton of quick illustrations or AI-generated images for pretty cheap, I think getting a hundred example sentences illustrated might be outside my budget for the project.
@pganssle Ah wow, awesome use case. My first job out of uni was at Memrise, which had the concept of “mems” - user generated reminder images based on some kind of association. Love to see the idea still being used. I’m a big Anki user too 😁
And yeah, AI generated art isn’t exactly fast when you have some idea…
@adamchainz I kinda wonder if the early "hands with too many fingers" and bizarre extra limbs and stuff made people try to put huge penalties in for bizarre stuff like that, and it gets penalized when you actually *try* to generate something weird.