Loss-aware Binarization of Deep Networks

Lu Hou, Quanming Yao, James T. Kwok

Deep neural network models, though very powerful and highly successful, are computationally expensive in terms of space and time. Recently, there have been a number of attempts on…