3 ms·
The point of the original paper was to learn the hyperparameters of a DNN using a DNN, as opposed to using, say, a bayesian optimization framework.
by dougabug 10y ago
The point of the original paper was to learn the hyperparameters of a DNN using a DNN, as opposed to using, say, a bayesian optimization framework.