CondenseNet: An Efficient DenseNet using Learned Group Convolutions

TL;DR

CondenseNet introduces a highly efficient neural network architecture that combines dense connectivity with learned group convolutions, significantly outperforming existing compact models in resource-constrained environments.

Contribution

The paper presents CondenseNet, a novel architecture that integrates learned group convolutions with dense connectivity for improved efficiency and feature reuse.

Findings

CondenseNet outperforms MobileNets and ShuffleNets in efficiency.

Learned group convolutions reduce unnecessary connections.

Dense connectivity enhances feature reuse.

Abstract

Deep neural networks are increasingly used on mobile devices, where computational resources are limited. In this paper we develop CondenseNet, a novel network architecture with unprecedented efficiency. It combines dense connectivity with a novel module called learned group convolution. The dense connectivity facilitates feature re-use in the network, whereas learned group convolutions remove connections between layers for which this feature re-use is superfluous. At test time, our model can be implemented using standard group convolutions, allowing for efficient computation in practice. Our experiments show that CondenseNets are far more efficient than state-of-the-art compact convolutional networks such as MobileNets and ShuffleNets.