1 paper
Yuzhang Shang, Dan Xu, Ziliang Zong +2
Neural network binarization accelerates deep models by quantizing their weights and activations into 1-bit. However, there is still a huge performance gap between Binary Neural Net…