Lipschitz Continuity Retained Binary Neural Network

Yuzhang Shang, Dan Xu, Bin Duan +3

Relying on the premise that the performance of a binary neural network can be largely restored with eliminated quantization error between full-precision weight vectors and their co…