BNN

Binarized Neural Networks NIPS2016

$$ \hat{w}_q = \mbox{sign}(\hat{w}) $$

Experiments

exp

Reference:

[1] Binaryconnect: Training deep neural networks with binary weights during propagations.