English

Resolution Switchable Networks for Runtime Efficient Image Recognition

Computer Vision and Pattern Recognition 2020-11-10 v3

Abstract

We propose a general method to train a single convolutional neural network which is capable of switching image resolutions at inference. Thus the running speed can be selected to meet various computational resource limits. Networks trained with the proposed method are named Resolution Switchable Networks (RS-Nets). The basic training framework shares network parameters for handling images which differ in resolution, yet keeps separate batch normalization layers. Though it is parameter-efficient in design, it leads to inconsistent accuracy variations at different resolutions, for which we provide a detailed analysis from the aspect of the train-test recognition discrepancy. A multi-resolution ensemble distillation is further designed, where a teacher is learnt on the fly as a weighted ensemble over resolutions. Thanks to the ensemble and knowledge distillation, RS-Nets enjoy accuracy improvements at a wide range of resolutions compared with individually trained models. Extensive experiments on the ImageNet dataset are provided, and we additionally consider quantization problems. Code and models are available at https://github.com/yikaiw/RS-Nets.

Keywords

Cite

@article{arxiv.2007.09558,
  title  = {Resolution Switchable Networks for Runtime Efficient Image Recognition},
  author = {Yikai Wang and Fuchun Sun and Duo Li and Anbang Yao},
  journal= {arXiv preprint arXiv:2007.09558},
  year   = {2020}
}

Comments

ECCV 2020. Code and models: https://github.com/yikaiw/RS-Nets

R2 v1 2026-06-23T17:13:20.765Z