arXiv 1804.032302018-04cs.CV

NetAdapt：面向移动应用的平台感知神经网络适配

NetAdapt: Platform-Aware Neural Network Adaptation for Mobile Applications

Massachusetts Institute of Technology（麻省理工学院）；Google Inc.（谷歌公司）

Tien-Ju Yang, Andrew Howard, Bo Chen, Xiao Zhang, Alec Go, Mark Sandler, Vivienne Sze, Hartwig Adam

备注：: Accepted by ECCV 2018

摘要：

这项工作提出了一种名为NetAdapt的算法，该算法可以在给定资源预算的情况下自动将预先训练的深度神经网络适应移动平台。虽然许多现有算法基于MAC或权重的数量来简化网络，但优化这些间接指标可能不一定会减少直接指标，例如延迟和能耗。为了解决这个问题，NetAdapt将直接度量纳入其自适应算法。这些直接指标使用经验测量进行评估，因此不需要平台和工具链的详细知识。NetAdapt自动并逐步简化预先训练的网络，直到满足资源预算，同时最大限度地提高准确性。实验结果表明，与现有的自动化网络简化算法相比，NetAdapt在移动CPU和移动GPU上都取得了更好的精度和延迟平衡。对于ImageNet数据集上的图像分类，NetAdapt在测量的推理延迟方面实现了高达1.7 $\times$的加速，在MobileNets（V1和V2）上具有相同或更高的准确性。

英文摘要：

This work proposes an algorithm, called NetAdapt, that automatically adapts a pre-trained deep neural network to a mobile platform given a resource budget. While many existing algorithms simplify networks based on the number of MACs or weights, optimizing those indirect metrics may not necessarily reduce the direct metrics, such as latency and energy consumption. To solve this problem, NetAdapt incorporates direct metrics into its adaptation algorithm. These direct metrics are evaluated using empirical measurements, so that detailed knowledge of the platform and toolchain is not required. NetAdapt automatically and progressively simplifies a pre-trained network until the resource budget is met while maximizing the accuracy. Experiment results show that NetAdapt achieves better accuracy versus latency trade-offs on both mobile CPU and mobile GPU, compared with the state-of-the-art automated network simplification algorithms. For image classification on the ImageNet dataset, NetAdapt achieves up to a 1.7$\times$ speedup in measured inference latency with equal or higher accuracy on MobileNets (V1&V2).

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。

摘要：

英文摘要：

相关论文