Compressing deep quaternion neural networks with targeted regularization

Open in new window