Group-Shared Low-Rank Approximation for Mobile-Efficient Pointwise Convolutions in Large-Kernel CNNs
2026-08-26 • Machine Learning
Machine Learning
AI summaryⓘ
The authors found that large-kernel CNNs work well but are too big to easily run on devices like smartphones because the pointwise convolutions use most of the storage. To fix this, they created a new method called Channel Group-Shared (CGS) low-rank approximation that shares parameters cleverly across channel groups, greatly shrinking the model size. Their experiments show that CGS keeps good performance while making the models much smaller and faster to load. This makes it easier to use powerful vision CNNs on devices with limited memory and storage.
Large-kernel CNNPointwise convolutionDepthwise separable convolutionLow-rank approximationSingular Value Decomposition (SVD)Parameter sharingMemory bandwidthEdge devicesModel compression
Authors
Hao Luo, Yiting Yang, Wenyi Zhao, Man Jiang, Zhijun Lin, Ghulam Mohiuddin, Ting Jiang, Kunming Luo, Zihao Zhang, Qingsen Yan, Guoqing Wang, Wei Dong, Peng Wang
Abstract
Large-kernel Convolutional Neural Networks (CNNs) deliver remarkable performance in vision tasks by significantly expanding receptive fields, yet their quadratic parameter growth critically impedes storage-efficient edge deployment. While existing efficient architectures adopt parameter-efficient depthwise separable convolution backbones that leverage techniques like low-rank approximation and weight sharing to compress depthwise convolutions, we identify a critical oversight: pointwise convolutions dominate parameter volume (>87% in models like RepLKNet-31B) and constitute the primary deployment bottleneck on resource-constrained edge devices. This results in prohibitive storage costs and severe memory-loading constraints on resource-limited devices (e.g., smartphones with 4-12 GB Random Access Memory (RAM)). To overcome this, we propose Channel Group-Shared (CGS) low-rank approximation, a novel Singular Value Decomposition (SVD)-based parameter-sharing strategy. CGS constructs a structured low-rank paradigm isomorphic to SVD decomposition, comprising shared (high-parameter-cost) down/up-projection matrices across channel groups within a layer and channel-group-specific (low-parameter-cost) scalable diagonal matrices. This group-sharing design achieves significant parameter reduction. Extensive experiments demonstrate that large-kernel CNNs (RepLKNet, ConvNeXt, SLaK) enhanced with CGS strike an empirically favorable balance between competitive performance and substantially reduced storage costs. Crucially, by alleviating storage constraints, reducing memory bandwidth pressure during loading, and minimizing model loading latency, CGS enables the feasible deployment of pre-trained large-kernel CNN models on edge devices, thereby bridging the gap between high-performance vision models and practical edge deployment.