AIoTBench
收藏arXiv2020-05-07 更新2024-08-06 收录
下载链接:
http://arxiv.org/abs/2005.05085v1
下载链接
链接失效反馈官方服务:
资源简介:
AIoTBench是由中国科学院计算技术研究所开发的基准套件,专注于移动和嵌入式设备的推理能力评估。该数据集包含5000张从ImageNet 2012验证集中随机抽样的图像,用于评估包括ResNet50、InceptionV3等在内的多种网络模型。AIoTBench旨在通过统一的VIPS和VOPS指标,量化不同设备在处理图像分类任务时的性能,从而为移动AI应用提供全面的性能评估和比较。
AIoTBench is a benchmark suite developed by the Institute of Computing Technology, Chinese Academy of Sciences, focusing on evaluating the inference capabilities of mobile and embedded devices. This dataset contains 5,000 images randomly sampled from the ImageNet 2012 validation set, and is utilized to evaluate multiple network models including ResNet50, InceptionV3 and others. AIoTBench aims to quantify the performance of various devices when processing image classification tasks through unified VIPS and VOPS metrics, thereby providing comprehensive performance evaluation and comparison for mobile AI applications.
提供机构:
中国科学院计算技术研究所
创建时间:
2020-05-07



