chentastic 9 days ago

Was always fascinated by SVG art. How good are LLMs in generating SVGs?

6
jamesgpearce 9 days ago

Perhaps a generous definition of "better"... but some of those genuinely made me laugh!

simpaticoder 9 days ago

Regular LLMs are quite bad at it (see simonwillison's blog post). However this paper [0] describes an apparently sound approach using Neural Radiance Fields (NeRFs), however their github repo [1] has been "code coming soon!" for months now, so you can't really use it.

0 - https://arxiv.org/pdf/2501.03992

1 - https://github.com/SagiPolaczek/NeuralSVG

flaviuspopan 4 days ago

It's been released as of an hour ago!!! https://github.com/SagiPolaczek/NeuralSVG/pull/6

Oh boy oh boy oh boy

plumeria 9 days ago
pizza 9 days ago

I've gotten decent outputs with Claude with iteration (sending both text feedback and screenshot for context) and then tweaked the output in Inkscape.

jbreckmckye 9 days ago

In at least my limited experience, they're kind of bad. They can retrieve shapes that already exist, sometimes inaccurately, but they are less reliable at creating novel ones