arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2101.07985cs.CV
收藏

使用自适应示例滤波器的网络修剪

Network Pruning using Adaptive Exemplar Filters

Xiamen University(厦门大学);Pinterest(Pinterest(拼趣));Tencent(腾讯);University of Chinese Academy of Sciences(中国科学院大学);Peng Cheng Lab(鹏城实验室)

Mingbao Lin, Rongrong Ji, Shaojie Li, Yan Wang, Yongjian Wu, Feiyue Huang, Qixiang Ye

摘要:

流行的网络剪枝算法通过优化手工构建的模型来减少冗余信息,可能会导致性能不佳和选择过滤器的时间过长。我们创新性地引入自适应样本过滤器来简化算法设计,从而产生一种自动且高效的剪枝方法,称为EPruner。受人脸识别社区的启发,我们在权重矩阵上使用消息传递算法Affinity Propagation来获得自适应数量的样本,然后将其作为保留的过滤器。EPruner在确定“重要”过滤器时打破了对训练数据的依赖,并允许在几秒钟内实现CPU,比基于GPU的SOTA快一个数量级。此外,我们表明,样本的权重提供了一个更好的初始化微调。在VGGNet-16上,EPruner通过删除88.80%的参数实现了76.34%的FLOP减少,在CIFAR-10上提高了0.06%的准确性。在ResNet-152中,EPruner通过删除64.18%的参数实现了65.12%的FLOP减少,在ILSVRC-2012上仅损失了0.71%的前5个精度。我们的代码可以在https://github.com/lmbxmu/EPruner上获得。

英文摘要:

Popular network pruning algorithms reduce redundant information by optimizing hand-crafted models, and may cause suboptimal performance and long time in selecting filters. We innovatively introduce adaptive exemplar filters to simplify the algorithm design, resulting in an automatic and efficient pruning approach called EPruner. Inspired by the face recognition community, we use a message passing algorithm Affinity Propagation on the weight matrices to obtain an adaptive number of exemplars, which then act as the preserved filters. EPruner breaks the dependency on the training data in determining the "important" filters and allows the CPU implementation in seconds, an order of magnitude faster than GPU based SOTAs. Moreover, we show that the weights of exemplars provide a better initialization for the fine-tuning. On VGGNet-16, EPruner achieves a 76.34%-FLOPs reduction by removing 88.80% parameters, with 0.06% accuracy improvement on CIFAR-10. In ResNet-152, EPruner achieves a 65.12%-FLOPs reduction by removing 64.18% parameters, with only 0.71% top-5 accuracy loss on ILSVRC-2012. Our code can be available at https://github.com/lmbxmu/EPruner.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑