←back to thread

Diffusion training from scratch on a micro-budget

(github.com)
135 points lnyan | 1 comments | 13 Jan 25 09:55 UTC | HN request time: 0.2s | source
Show context
__loam ◴[16 Jan 25 22:33 UTC] No.42731765[source]▶
The pixel art these models produce continues to look like shit and not be actual pixel art.

Where'd you get your dataset? Did you get permission from the rightsholders to use their work for this?

replies(1): >>42732011 #
GaggiX ◴[16 Jan 25 23:00 UTC] No.42732011[source]▶
This model is kind of the bottom of the barrel. Also there are models that specialize in pixel art: https://www.pixellab.ai/
replies(1): >>42733893 #
1. AuryGlenz ◴[17 Jan 25 03:53 UTC] No.42733893[source]▶
I’m guessing that’s built on Stable Diffusion/Flux just as RetroDiffusion is. That would mean it’s not directly pixel art and needs to be downscaled after the fact. The results can still be pretty great but it’s not super ideal.

Training a small (64x64 or the like) pixel based model instead of something that relies on a VAE may in fact be cheaper than what’s in the OP and could probably make someone some good money besides but the lack of training data would be a huge issue.