by tikwidd on 8/31/22, 2:29 AM with 4 comments
What would it take to get image synthesis and img2img for generating SVG graphics?
by minimaxir on 8/31/22, 3:12 AM
It's probably easier/faster to just have AI generate vector art then use a tool to automatically vectorize it.
by sonovice on 9/1/22, 7:43 AM
by dougabug on 8/31/22, 2:41 AM
Diffusion models can be trained to learn distributions over more or less arbitrary spaces of signals.