> Training a one bit neural network from scratch is apparently an unsolved problem though. It was until recently, but there is a new method which trains them d

> Training a one bit neural network from scratch is apparently an unsolved problem though. It was until recently, but there is a new method which trains them directly without any floating point math, using "Boolean variation" instead of Newton/Leibniz differentiation: https://proceedings.neurips.cc/paper_files/paper/2024/hash/718a3c5cf135894db6e718725f52ef9a-Abstract-Conference.html https://proceedings.neurips.cc/paper_files/paper/2024/hash/7...