Accelerating Very Deep Convolutional Networks for Classification and Detection

TL;DR

This paper presents a nonlinear optimization method to accelerate very deep CNNs, achieving a 4x speedup on VGG-16 with minimal accuracy loss, applicable to classification and detection tasks.

Contribution

It introduces a nonlinear, layer-wise approximation technique that reduces accumulated errors in deep CNNs without using SGD, improving speed while maintaining accuracy.

Findings

Achieved 4x speedup on VGG-16 with only 0.3% top-5 error increase.

Enabled effective acceleration for deep CNNs in classification and detection.

Reduced error accumulation across multiple layers in deep networks.

Abstract

This paper aims to accelerate the test-time computation of convolutional neural networks (CNNs), especially very deep CNNs that have substantially impacted the computer vision community. Unlike previous methods that are designed for approximating linear filters or linear responses, our method takes the nonlinear units into account. We develop an effective solution to the resulting nonlinear optimization problem without the need of stochastic gradient descent (SGD). More importantly, while previous methods mainly focus on optimizing one or two layers, our nonlinear method enables an asymmetric reconstruction that reduces the rapidly accumulated error when multiple (e.g., >=10) layers are approximated. For the widely used very deep VGG-16 model, our method achieves a whole-model speedup of 4x with merely a 0.3% increase of top-5 error in ImageNet classification. Our 4x accelerated VGG-16 model also shows a graceful accuracy degradation for object detection when plugged into the Fast R-CNN detector.