Approach

The new architecture utilizes two proposed operations, pointwise group convolution and channel shuffle, to greatly reduce computation cost while maintaining accuracy.

We notice 1 × 1 convolutions require(also called pointwise convolutions) considerable complexity. To address the issue, a straightforward solution is to apply channel sparse connections, for example group convolutions, also on 1 × 1 layers. However, this property blocks information flow between channel groups and weakens representation.

If we allow 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.This can be efficiently and elegantly implemented by a channel shuffle operation.

• ShuffleNet Unit

• Network Architecture

Experiment

References：
ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices， Xiangyu Zhang