Building Efficient ConvNets using Redundant Feature Pruning

Babajide O. Ayinde, Jacek M. Zurada
2/21/2018
cs.CV

Official code: babajide07/Redundant-Feature-Pruning-Pytorch-Implementation (16 stars, last commit 2020-04-20).

Abstract

This paper presents an efficient technique to prune deep and/or wide convolutional neural network models by eliminating redundant features (or filters). Previous studies have shown that over-sized deep neural network models tend to produce a lot of redundant features that are either shifted version of one another or are very similar and show little or no variations; thus resulting in filtering redundancy. We propose to prune these redundant features along with their connecting feature maps according to their differentiation and based on their relative cosine distances in the feature space, thus yielding smaller network size with reduced inference costs and competitive performance. We empirically show on select models and CIFAR-10 dataset that inference costs can be reduced by 40% for VGG-16, 27% for ResNet-56, and 39% for ResNet-110.

View on arXivView PDF

Code Implementations(1)

babajide07/Redundant-Feature-Pruning-Pytorch-ImplementationOfficial100%
162PythonDec 29, 20176 years ago

Similar Papers

Data-Driven Sparse Structure Selection for Deep Neural Networks

87% match1 repo392017

Learning Efficient Convolutional Networks through Network Slimming

85% match2 repos9192017

Dynamic Channel Pruning: Feature Boosting and Suppression

85% match1 repo1122018

Structured Network Pruning by Measuring Filter-wise Interactions

85% match2023

HRank: Filter Pruning using High-Rank Feature Map

85% match1 repo2562020

Visual Prompting Upgrades Neural Network Sparsification: A Data-Model Perspective

85% match1 repo122023

Cite this paper

@article{ayinde2018building,
  title  = {Building Efficient ConvNets using Redundant Feature Pruning},
  author = {Babajide O. Ayinde and Jacek M. Zurada},
  year   = {2018},
  eprint = {1802.07653},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1802.07653}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1802.07653.svg)](https://codeofpaper.com/paper/1802.07653)

Discussion