/principled-dfa-training

Code for our paper on best practices to train neural networks with direct feedback alignment (DFA).

Primary LanguagePython

Watchers