# Fast Training of Convolutional Neural Networks via Kernel Rescaling

Source: [Fast Training of Convolutional Neural Networks via Kernel Rescaling](<https://www.sota2.com/research/paper/68fa62ce1e209131ce9fb286>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T20:26:18.516Z

Authors: Pedro Porto Buarque de Gusm\\~ao, Gianluca Francini, Skjalg Leps{\\o}y, Enrico Magli

Published: 2016-10-12T07:25:34.000Z

## Abstract

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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1610.03623>)

[PDF](<https://arxiv.org/pdf/1610.03623.pdf>)