4 ms·Incorporating Prior Information to Speed Up Neural Network Training by 3.5X2 points by jtsymonds 8y ago