Fetching the paper…

On the Importance of Normalisation Layers in Deep Learning with Piecewise Linear Activation Units · Around