Fetching the paper…

GPU Asynchronous Stochastic Gradient Descent to Speed Up Neural Network Training · Around