Tune smarter not harder: A principled approach to tuning learning rates for shallow nets

Open in new window