DOOMFORMER
github.com/mottopanikeiku/doomformerAbout
DOOMFORMER compiles integer neural networks into DOOM maps, where voodoo dolls perform the arithmetic and light levels and plane heights store the values. Results are read back from ViZDoom and checked against a NumPy integer reference with zero tolerance. The README reports exact classification of 10,000 MNIST test digits, and demos include a marine that is steered by the in-map network.
Tags
doomvizdoomneural-networkstransformeresotericpython
Tech Stack
Python
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.