Fetching the paper…

On the Computational Inefficiency of Large Batch Sizes for Stochastic Gradient Descent · Around