2021/04/24 by Ting-Wu Chin, Chin, Ting-Wu, Diana Marculescu +3
Computer Science · #Advanced Neural Network Applications #Adversarial Robustness in Machine Learning #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning (cs.LG) #cs.CV #cs.LG
paper · pdf · doi:10.48550/arxiv.2104.13255
Full paper accepted at CVPR Workshops 2021; a 4-page abridged version is accepted at ICLR 2021 NAS Workshop
arxiv created 2021/04/24 · openalex publication_date 2021/04/24 · arxiv updated 2021/04/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Optimizing the channel counts for different layers of a CNN has shown great promise in improving the efficiency of CNNs at test-time. However, these methods often introduce large computational overhead (e.g., an additional 2x FLOPs of standard training). Minimizing this overhead could therefore significantly speed up training. In this work, we propose width transfer, a technique that harnesses the assumptions that the optimized widths (or channel counts) are regular across sizes and depths. We show that width transfer works well across various width optimization algorithms and networks. Specifically, we can achieve up to 320x reduction in width optimization overhead without compromising the top-1 accuracy on ImageNet, making the additional cost of width optimization negligible relative to initial training. Our findings not only suggest an efficient way to conduct width optimization but also highlight that the widths that lead to better accuracy are invariant to various aspects of network architectures and training data.