Soft independence guided filter pruning

Open Webpage

Liu Yang, Shiqiao Gu, Chenyang Shen, Xile Zhao, Qinghua Hu

Published: 01 Jan 2024, Last Modified: 04 Mar 2025Pattern Recognit. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Filter pruning considers both the magnitude and correlation information.•Cross-filter-independence-based metric increases efficiency and lowers costs.•Asymptotic pruning ratio progressively concentrates training information.