Positional Normalization

          Boyi Li1,2, Felix Wu1, Kilian Q. Weinberger1 , Serge Belongie1,2
                   1
                    Cornell University 2 Cornell Tech
           {bl728, fw245, kilian, sjb344}@cornell.edu

                         Abstract
       A popular method to reduce the training time of deep neural networks is to normal-
       ize activations at each layer. Although various normalization schemes have been
       proposed, they all follow a common theme: normali ...