Finet: Using Fine-grained Batch Normalization to Train Light-weight Neural Networks

Open in new window