Training Deep Neural Networks by optimizing over nonlocal paths in hyperparameter space

Open in new window