剪枝主胶囊让胶囊网络提速近10倍，精度不降
PrunedCaps: A Case For Primary Capsules Discrimination
- 通过剪枝移除95%主胶囊，大幅降低计算量
- 在多个数据集上精度不变，推理速度提升9.9倍
- 揭示不同数据集对剪枝的敏感性差异
胶囊网络（CapsNets）相较于卷积神经网络（CNNs）在仿射变换鲁棒性和重叠图像检测方面具有优势。然而，由于主胶囊（PCs）数量过多，其训练与测试过程缓慢且资源消耗大，难以被视为高效架构。本文在手写数字MNIST、Fashion-MNIST、CIFAR-10和SVHN数据集上研究了主胶囊剪枝的可行性。结果表明，剪枝后的胶囊网络在移除95%胶囊的情况下，仍保持原有精度，推理速度提升达9.90倍；同时，在动态路由阶段节省超过95.36%的浮点运算量。此外，论文分析了为何某些数据集剪枝后表现显著提升，而另一些则效果不佳。
原文摘要 · Abstract (English)
Capsule Networks (CapsNets) are a generation of image classifiers with proven advantages over Convolutional Neural Networks (CNNs). Better robustness to affine transformation and overlapping image detection are some of the benefits associated with CapsNets. However, CapsNets cannot be classified as resource-efficient deep learning architecture due to the high number of Primary Capsules (PCs). In addition, CapsNets' training and testing are slow and resource hungry. This paper investigates the possibility of Primary Capsules pruning in CapsNets on MNIST handwritten digits, Fashion-MNIST, CIFAR-10, and SVHN datasets. We show that a pruned version of CapsNet performs up to 9.90 times faster than the conventional architecture by removing 95 percent of Capsules without a loss of accuracy. Also, our pruned architecture saves on more than 95.36 percent of floating-point operations in the dynamic routing stage of the architecture. Moreover, we provide insight into why some datasets benefit significantly from pruning while others fall behind.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。