优化克利福德神经层推理速度，实测比PyTorch快30%
Fast Clifford Neural Layers
- 基于克利福德代数设计高效卷积与多向量激活层
- 在大尺寸数据下实现比PyTorch快30%的推理性能
- 适合需要高性能物理方程建模的科研与工程应用
克利福德神经层通过将克利福德代数引入神经网络，提升了偏微分方程（PDE）建模能力。本项目聚焦于优化2/3D克利福德卷积层与多向量激活层在单核CPU上的推理性能。在包含克利福德卷积层和多向量激活层的真实网络块测试中，当数据与网络规模超过L2缓存时，我们的实现比标准PyTorch版本快30%。代码已开源，地址为https://github.com/egretwAlker/c-opt-clifford-layers。
原文摘要 · Abstract (English)
Clifford Neural Layers improve PDE modeling by introducing Clifford Algebra into neural networks. In this project we focus on optimizing the inference of 2/3D Clifford convolutional layers and multivector activation layers for one core CPU performance. Overall, by testing on a real network block involving Clifford convolutional layers and multivector activation layers, we observe that our implementation is 30% faster than standard PyTorch implementation in relatively large data + network size (>L2 cache). We open source our code base at https://github.com/egretwAlker/c-opt-clifford-layers
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。