MikeTrendsTrends right now

⬢github Python · 57 ★ +55 since we first saw it · pushed 2 h ago · MIT

mottopanikeiku/doomformer

Compile neural networks into DOOM maps, where voodoo dolls do the arithmetic

DOOMFORMER implements neural networks purely with DOOM's game mechanics: numbers are crowds of voodoo dolls on conveyor belts, matrix multiplies cross lava lines, and attention squeezes under doors. All computed values are verified bit-exact against a NumPy integer reference, and demos include an MNIST digit reader played inside the game and a transformer built into a DOOM level.

Why now: Featured on Hacker News ('I built a neural network using DOOM games'), a novel esoteric-computing demo with watchable films and a Colab notebook.

Who it is for: Curious developers and esoteric-computing fans who enjoy unconventional computing demos in DOOM.

doomesoteric-computingneural-networksmnistpythontransformer

compilerdoomesoteric-computingmnistneural-networkstransformervizdoomvoodoo-dolls

Open on GitHub →

Stars over our 799 snapshots: 2 to 57, since 9 d ago.

Where people talked about it

API: https://socialmediatrends-api.osmike.com/v1/repos/mottopanikeiku/doomformer