Automatic Gradient Descent: Deep Learning without Hyperparameters

Open in new window