Channel Shuffle is an operation to help information flow across feature channels in convolutional neural networks. It was used as part of the ShuffleNet architecture.

If we allow a group convolution to obtain input data from different groups, the input and output channels will be fully related. Specifically, for the feature map generated from the previous group layer, we can first divide the channels in each group into several subgroups, then feed each group in the next layer with different subgroups.

The above can be efficiently and elegantly implemented by a channel shuffle operation: suppose a convolutional layer with $g$ groups whose output has $g \times n$ channels; we first reshape the output channel dimension into $\left(g, n\right)$, transposing and then flattening it back as the input of next layer. Channel shuffle is also differentiable, which means it can be embedded into network structures for end-to-end training.

Source: ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices

Latest Papers

PAPER DATE
MimicNorm: Weight Mean and Last BN Layer Mimic the Dynamic of Batch Normalization
| Wen FeiWenrui DaChenglin LiJunni ZouHongkai Xiong
2020-10-19
Rethinking Image Deraining via Rain Streaks and Vapors
Yinglong WangYibing SongChao MaBing Zeng
2020-08-03
Egocentric Human Segmentation for Mixed Reality
Andrija GajicEster Gonzalez-SosaDiego Gonzalez-MorinMarcos Escudero-ViñoloAlvaro Villegas
2020-05-25
Kaleidoscope: An Efficient, Learnable Representation For All Structured Linear Maps
| Tri DaoNimit SohoniAlbert GuMatthew EichhornAmit BlonderMegan LeszczynskiAtri RudraChristopher Ré
2020-05-01
DyNet: Dynamic Convolution for Accelerating Convolutional Neural Networks
Yikang ZhangJian ZhangQiang WangZhao Zhong
2020-04-22
Model-Agnostic Structured Sparsification with Learnable Channel Shuffle
Xin-Yu ZhangKai ZhaoTaihong XiaoMing-Ming ChengMing-Hsuan Yang
2020-02-19
Precision Gating: Improving Neural Network Efficiency with Dynamic Dual-Precision Activations
| Yichi ZhangRitchie ZhaoWeizhe HuaNayun XuG. Edward SuhZhiru Zhang
2020-02-17
Pre-defined Sparsity for Low-Complexity Convolutional Neural Networks
Souvik KunduMahdi NazemiMassoud PedramKeith M. ChuggPeter A. Beerel
2020-01-29
CNN-CASS: CNN for Classification of Coronary Artery Stenosis Score in MPR Images
Mariia DobkoBohdan PetryshakOles Dobosevych
2020-01-23
Quantisation and Pruning for Neural Network Compression and Regularisation
| Kimessha PaupamahSteven JamesRichard Klein
2020-01-14
Kaleidoscope: An Efficient, Learnable Representation For All Structured Linear Maps
| Anonymous
2020-01-01
DyNet: Dynamic Convolution for Accelerating Convolution Neural Networks
Anonymous
2020-01-01
Resizable Neural Networks
Anonymous
2020-01-01
ZeroQ: A Novel Zero Shot Quantization Framework
| Yaohui CaiZhewei YaoZhen DongAmir GholamiMichael W. MahoneyKurt Keutzer
2020-01-01
LiteSeg: A Novel Lightweight ConvNet for Semantic Segmentation
| Taha EmaraHossam E. Abd El MunimHazem M. Abbas
2019-12-13
Depth-wise Decomposition for Accelerating Separable Convolutions in Efficient Convolutional Neural Networks
Yihui HeJianing QianJianren Wang
2019-10-21
A Pre-defined Sparse Kernel Based Convolution for Deep CNNs
Souvik KunduSaurav PrakashHaleh AkramiPeter A. BeerelKeith M. Chugg
2019-10-02
ThunderNet: Towards Real-Time Generic Object Detection on Mobile Devices
Zheng Qin Zeming Li Zhaoning Zhang Yiping Bao Gang Yu Yuxing Peng Jian Sun
2019-10-01
EdgeCNN: Convolutional Neural Network Classification Model with small inputs for Edge Computing
| Shunzhi YangZheng GongKai YeYungen WeiZheng HuangZhenhua Huang
2019-09-30
Go Wider: An Efficient Neural Network for Point Cloud Analysis via Group Convolutions
Can ChenLuca Zanotti FragonaraAntonios Tsourdos
2019-09-23
Mish: A Self Regularized Non-Monotonic Neural Activation Function
| Diganta Misra
2019-08-23
RRNet: Repetition-Reduction Network for Energy Efficient Decoder of Depth Estimation
Sangyun OhHye-Jin S. KimJongeun LeeJunmo Kim
2019-07-23
Deep Single Image Deraining Via Estimating Transmission and Atmospheric Light in rainy Scenes
Yinglong WangQinfeng ShiEhsan AbbasnejadChao MaXiaoping MaBing Zeng
2019-06-22
HGC: Hierarchical Group Convolution for Highly Efficient Neural Network
Xukai XieYuan ZhouSun-Yuan Kung
2019-06-09
DiCENet: Dimension-wise Convolutions for Efficient Networks
| Sachin MehtaHannaneh HajishirziMohammad Rastegari
2019-06-08
Butterfly Transform: An Efficient FFT Based Neural Architecture Design
| Keivan Alizadeh VahidAnish PrabhuAli FarhadiMohammad Rastegari
2019-06-05
Multi-Person Pose Estimation with Enhanced Channel-wise and Spatial Information
Kai SuDongdong YuZhenqi XuXin GengChanghu Wang
2019-05-09
ThunderNet: Towards Real-time Generic Object Detection
| Zheng QinZeming LiZhaoning ZhangYiping BaoGang YuYuxing PengJian Sun
2019-03-28
DetNAS: Backbone Search for Object Detection
| Yukang ChenTong YangXiangyu ZhangGaofeng MengXinyu XiaoJian Sun
2019-03-26
Learnable Embedding Space for Efficient Neural Architecture Compression
| Shengcao CaoXiaofang WangKris M. Kitani
2019-02-01
Slimmable Neural Networks
| Jiahui YuLinjie YangNing XuJianchao YangThomas Huang
2018-12-21
ShuffleNASNets: Efficient CNN models through modified Efficient Neural Architecture Search
Kevin Alexander LaubeAndreas Zell
2018-12-07
DSNet for Real-Time Driving Scene Semantic Segmentation
Wenfu WangZhijie Pan
2018-12-06
Shift-based Primitives for Efficient Convolutional Neural Networks
Huasong ZhongXianggen LiuYihui HeYuchun Ma
2018-09-22
MBS: Macroblock Scaling for CNN Model Reduction
Yu-Hsun LinChun-Nan ChouEdward Y. Chang
2018-09-18
ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
| Ningning MaXiangyu ZhangHai-Tao ZhengJian Sun
2018-07-30
Merging and Evolution: Improving Convolutional Neural Networks for Mobile Applications
| Zheng QinZhaoning ZhangShiqing ZhangHao YuYuxing Peng
2018-03-24
RTSeg: Real-time Semantic Segmentation Comparative Study
| Mennatullah SiamMostafa GamalMoemen Abdel-RazekSenthil YogamaniMartin Jagersand
2018-03-07
HENet:A Highly Efficient Convolutional Neural Networks Optimized for Accuracy, Speed and Storage
| Qiuyu ZhuRuixin Zhang
2018-03-07
FD-MobileNet: Improved MobileNet with a Fast Downsampling Strategy
| Zheng QinZhaoning ZhangXiaotao ChenYuxing Peng
2018-02-11
EffNet: An Efficient Structure for Convolutional Neural Networks
| Ido FreemanLutz Roese-KoernerAnton Kummert
2018-01-19
ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices
| Xiangyu ZhangXinyu ZhouMengxiao LinJian Sun
2017-07-04

Components

COMPONENT TYPE
🤖 No Components Found You can add them if they exist; e.g. Mask R-CNN uses RoIAlign

Categories