Re:Show HN: Can an AI model fit on a single pixel?
by mememememememo on Wednesday April 08, 2026 @02:22AM
(#47686101)
It is fun. The trainer seems to get stuck in local minima though.
Re:Show HN: Can an AI model fit on a single pixel?
by deevelton on Wednesday April 08, 2026 @02:41AM
(#47686224)
Thanks; does this fix it? https://github.com/dvelton/ai-pixel/commit/ba04e1c649e0ec99b...
Re:Show HN: Can an AI model fit on a single pixel?
by mememememememo on Wednesday April 08, 2026 @03:42AM
(#47686737)
Tried again and no. To reproduce create a corridor e.g.
X O
X O
X O
X O
X O
Line can get jammed in non vertical positon.
Re:Show HN: Can an AI model fit on a single pixel?
by deevelton on Wednesday April 08, 2026 @08:52AM
(#47689513)
Increased default epochs to 2000 instead of 500, I think that fixed it but let me know if you still run into this.
Re:Show HN: Can an AI model fit on a single pixel?
by mememememememo on Thursday April 09, 2026 @03:47AM
(#47700500)
Works for me now, thanks!
Feed it two inputs (e.g. chance of rain and wind speed; this is one of the examples in the demo) and it learns to answer a yes/no question like "bring an umbrella?" It's a one-neuron binary classifier with three learned parameters: two weights and a bias. Those three numbers map directly to Red, Green, and Blue. Save the model: you get a 1x1 PNG. Load the pixel: you get your classifier back. The color is the model.