Fast Training of Convolutional Neural Networks via Kernel Rescaling

Pedro Porto Buarque de Gusmão, Gianluca Francini, Skjalg Lepsøy, Enrico Magli

Training deep Convolutional Neural Networks (CNN) is a time consuming task that may take weeks to complete. In this article we propose a novel, theoretically founded method for reducing CNN training time without incurring any loss in accuracy. The basic idea is to begin training with a pre-train network using lower-resolution kernels and input images, and then refine the results at the full resolution by exploiting the spatial scaling property of convolutions. We apply our method to the ImageNet winner OverFeat and to the more recent ResNet architecture and show a reduction in training time of nearly 20% while test set accuracy is preserved in both cases.

Knowledge Graph

arrow_drop_up

Comments

Sign up or login to leave a comment