· 2018
ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices
- 9,074citations
- 2018year
Short summary
ShuffleNet, a new CNN architecture, achieves 7.8% lower top-1 error than MobileNet on ImageNet classification with a 40 MFLOPs budget, and ~13x speedup over AlexNet on mobile devices.
AI-generated from the title and abstract; the full text is not read.
Key points
- Introduces ShuffleNet, a CNN architecture optimized for low-power mobile devices (10-150 MFLOPs).
- Utilizes pointwise group convolution and channel shuffle to reduce computation cost.
- Achieves 7.8% lower top-1 error than MobileNet on ImageNet classification at 40 MFLOPs.
- Demonstrates ~13x speedup over AlexNet on ARM mobile devices with comparable accuracy.
AI-generated from the title and abstract; the full text is not read.
Abstract
We introduce an extremely computation-efficient CNN architecture named ShuffleNet, which is designed specially for mobile devices with very limited computing power (e.g., 10-150 MFLOPs). The new architecture utilizes two new operations, pointwise group convolution and channel shuffle, to greatly reduce computation cost while maintaining accuracy. Experiments on ImageNet classification and MS COCO object detection demonstrate the superior performance of ShuffleNet over other structures, e.g. lower top-1 error (absolute 7.8%) than recent MobileNet [12] on ImageNet classification task, under the computation budget of 40 MFLOPs. On an ARM-based mobile device, ShuffleNet achieves ~13× actual speedup over AlexNet while maintaining comparable accuracy.
The authors' abstract, as published at the source., 2018 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionComputer Science