Neural architecture search methods seek optimal candidates with efficient weight-sharing supernet training. However, recent studies indicate poor ranking consistency about the performance between stand-alone architectures and shared-weight networks. In this paper, we present Prior-Guided One-shot NAS (PGONAS) to strengthen the ranking correlation of supernets. Specifically, we first explore the effect of activation functions and propose a balanced sampling strategy based on the Sandwich Rule to alleviate weight coupling in the supernet. Then, FLOPs and Zen-Score are adopted to guide the training of supernet with ranking correlation loss. Our PGONAS ranks 3rd place in the supernet Track Track of CVPR2022 Second lightweight NAS challenge. Code is available in https://github.com/pprp/CVPR2022-NAS?competition-Track1-3th-solution.
@article{arxiv.2206.13329,
title = {Prior-Guided One-shot Neural Architecture Search},
author = {Peijie Dong and Xin Niu and Lujun Li and Linzhen Xie and Wenbin Zou and Tian Ye and Zimian Wei and Hengyue Pan},
journal= {arXiv preprint arXiv:2206.13329},
year = {2022}
}
Comments
Official 3st Place Solution for the Second workshop Neural Architecture Search Second lightweight NAS Challenge 2022 - Track1 Supernet Track. Official leaderboard: https://aistudio.baidu.com/aistudio/competition/detail/149/0/leaderboard CVPR 2022 Workshop: https://cvpr-nas.com/competition