Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

NeuralSVG: An Implicit Representation for Text-to-Vector Generation

sagipolaczek.github.io

Post reply on HN
Print view View original

41–50 of 84 posts

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#41

Post by tasuki »

vipshek wrote: post #26

This is excellent! I think the utility of generating vectors is far, far greater than all the raster generation that's been a big focus thus far (DALL-E, Midjourney, etc). Those efforts have been incredibly impressive, of course, but raster outputs are so much more difficult to work with. You're forced to "upscale" or "inpaint" the rasters using subsequent generative AI calls to actually iterate towards something use…

Ah, we should be friends!

I'm not sure what else to add, except that these are exactly the thoughts I think, and it used to feel lonely ;)

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#43

Post by gcr »

toisanji wrote: post #42

This is a group applying vector generation to animations: https://www.youtube.com/@studyturtlehq The graphic fidelity has been slowly improving over time.

can you say more? all of these videos have less than 5 views and i can't find any explanation of their process

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#44

Post by jansan »

Kiro wrote: post #30
1970-01-01 wrote: post #27

Aside: I've been having a very hard time prompting ChatGPT to spit out ASCII art. It really seems to not be able to do it. Here is an ASCII art representation of a hopping rabbit: ``` (\(\ ( -.-) o_(")(") ``` This is a simple representation of a rabbit with its ears up and in a hopping stance. Let me know if you'd like me to adjust it!

Pretty good if you ask me. What would a proper hopping rabbit ASCII art look like?

Not sure, but that is a sitting rabbit.

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#45

Post by Lerc »

vipshek wrote: post #26

This is excellent! I think the utility of generating vectors is far, far greater than all the raster generation that's been a big focus thus far (DALL-E, Midjourney, etc). Those efforts have been incredibly impressive, of course, but raster outputs are so much more difficult to work with. You're forced to "upscale" or "inpaint" the rasters using subsequent generative AI calls to actually iterate towards something use…

There is also the possibility for using these images as guidance for rasterization models. Generate easily manipulatable and composible images as a first stage then add detail once the image composition is satisfactory.

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#46

Post by SillyUsername »

vipshek wrote: post #26

This is excellent! I think the utility of generating vectors is far, far greater than all the raster generation that's been a big focus thus far (DALL-E, Midjourney, etc). Those efforts have been incredibly impressive, of course, but raster outputs are so much more difficult to work with. You're forced to "upscale" or "inpaint" the rasters using subsequent generative AI calls to actually iterate towards something use…

My little project for the highly intricate, messy representation ;) https://github.com/KodeMunkie/shapesnap (it stands on the backs of giants, original was not mine). It's also available on npm.

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#47

Post by gwern »

vipshek wrote: post #26

This is excellent! I think the utility of generating vectors is far, far greater than all the raster generation that's been a big focus thus far (DALL-E, Midjourney, etc). Those efforts have been incredibly impressive, of course, but raster outputs are so much more difficult to work with. You're forced to "upscale" or "inpaint" the rasters using subsequent generative AI calls to actually iterate towards something use…

Have you looked at https://www.recraft.ai/ recently? The image quality of their vector outputs seems to have gotten quite good, although you obviously still wouldn't want to try to generate densely textured or photographic-like images like Midjourney excels at. (For https://gwern.net/dropcap last year or before, we had to settle for Midjourney and create a somewhat convoluted workflow through Recraft; but if I were making dropcaps now, I think the latest Recraft model would probably suffice.)

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#48

Post by andy_ppp »

I’m looking forward to seeing what this makes of Simon Willison’s LLM SVG generation test prompt: “Generate an SVG of a pelican riding a bicycle”.

It’s quite amazing the progress we are seeing in AI and it will keep getting better which is somewhat terrifying.

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#49

Post by nojvek »

andy_ppp wrote: post #48

I’m looking forward to seeing what this makes of Simon Willison’s LLM SVG generation test prompt: “Generate an SVG of a pelican riding a bicycle”. It’s quite amazing the progress we are seeing in AI and it will keep getting better which is somewhat terrifying.

I asked both Claude and ChatGPT o3 to "generate svg of mainland USA with black outline".

Tried various models and they got it hopelessly wrong. Claude does an okay job at "Generate an SVG of a pelican riding a bicycle"

Re: NeuralSVG: An Implicit Representation for Text-to-Vector Generation

#50

Post by datadrivenangel »

Lerc wrote: post #45
vipshek wrote: post #26

This is excellent! I think the utility of generating vectors is far, far greater than all the raster generation that's been a big focus thus far (DALL-E, Midjourney, etc). Those efforts have been incredibly impressive, of course, but raster outputs are so much more difficult to work with. You're forced to "upscale" or "inpaint" the rasters using subsequent generative AI calls to actually iterate towards something use…

There is also the possibility for using these images as guidance for rasterization models. Generate easily manipulatable and composible images as a first stage then add detail once the image composition is satisfactory.

Trivially possible with controlnets!
Post reply on HN