• The VOIDRUNNER'S CODEX is coming! Explore new worlds, fight oppressive empires, fend off fearsome aliens, and wield deadly psionics with this comprehensive boxed set expansion for 5E and A5E!

D&D General DALL·E 3 does amazing D&D art

Saracenus

Always In School Gamer
Merry Goth Christmas, can you tell who is the party necromancer?

1701734699321.png
 

log in or register to remove this ad

It's not as immediately obvious as the Empire waistline thing, but the AI is also really, really bad at saris. It knows there's a choli (a blouse) and that there is flowing fabric that drapes, but the saris it produces don't actually drape in a way that makes sense, and it's often hard to figure out exactly how it connects. Sometimes it just looks like big sleeves.

View attachment 336179View attachment 336180View attachment 336181View attachment 336182View attachment 336183View attachment 336184View attachment 336185View attachment 336186

Which makes sense - how does AI know what clothing is in the first place. I should stop trying to reproduce specific fashion, but I live in hope.

Actually, that the result of having, like D3, a multipurpose model. It won't understand clothing as you said, to the point that it's really unable to distinguish between clothing and people and how they interact. A generic model will understand that sari are to be drawn over a person, in the rough shape of a person, but it won't be able to understand that clothing drape a person. A sari-wearer might as well be a strange human with bizarre arms and legs as far as the model know. If you want to have more realistic saris, you'd need to use a more focused model, trained with many more images of sari so it will generate more "sari-like" saris. Which is not the purpose of D3 to be able to draw a lot of thing. There are other implementation that would allow it, to the point of having model that might have the drawback of being unable to draw any other clothes than sari... that wouldn't be good either.
 









Remove ads

Top