Comment on OmniSVG: A Unified Scalable Vector Graphics Generation Model

rizzothesmall@sh.itjust.works ⁨1⁩ ⁨week⁩ ago

• paraphrand@lemmy.world ⁨1⁩ ⁨week⁩ ago

  This is the first time I’ve seen a model target SVG drafting. Anything you have seen previously about unicorns or whatever was just someone experimenting with interesting edge case usage.

  Feeding a language model a bunch of vector art does not seem productive to me. So it makes sense that something like GPT4 sucks at it.

  source
• Even_Adder@lemmy.dbzer0.com ⁨1⁩ ⁨week⁩ ago

  It can do IMG to SVG. Check out the right side of this image:

  Image

  source

  • GenderNeutralBro@lemmy.sdf.org ⁨1⁩ ⁨week⁩ ago

    Hard to judge quality when what we’re seeing is practically a pixel-perfect recreation. The tricky part of automated vectorization is detecting and plotting curves in such a way that it scales correctly. Bad implementations will use too many elements, or include straight lines that should be parts of curves, etc. Those errors would not be visible in those low-res rasterizations.

    source

    • Even_Adder@lemmy.dbzer0.com ⁨1⁩ ⁨week⁩ ago

      The project page didn’t have a link to it, but there is a demo on HF.

      source

      • GenderNeutralBro@lemmy.sdf.org ⁨1⁩ ⁨week⁩ ago

        Just gave it a try. I couldn’t get coherent results from img-to-svg with a few different tests of low-res pixel art and high-res cartoons. txt-to-svg also gave me incoherent blobs even with simple prompts. Something must be wrong there. Is it working for anyone else?

        I might just try installing it locally when I get home.

        source