Improving the Trainability of Deep Neural Networks through Layerwise Batch-Entropy Regularization | Read Paper on Bytez