Group-Shared Low-Rank Approximation for Mobile-Efficient Pointwise Convolutions in Large-Kernel CNNs
用于大核卷积神经网络中移动高效逐点卷积的组共享低秩近似
机构 * Northwestern Polytechnical University(西北工业大学) ; Nanchang University(南昌大学) ; China Mobile Chengdu Institute of Research and Development(中国移动成都研究院) ; Hong Kong University of Science and Technology(香港科技大学) ; Institute of AI for Industries, Chinese Academy of Sciences(中国科学院人工智能产业研究院) ; University of Electronic Science and Technology of China(电子科技大学)
AI总结 针对大核CNN中逐点卷积占比过高导致边缘部署瓶颈的问题,提出通道组共享低秩近似方法,在保持性能的同时降低存储成本,实现大核CNN的边缘高效部署。
Comments 17 pages, 10 figures, accepted by MobiCom2026