arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.05080cs.LGcs.SD
收藏

从计算到消耗:探索用于 SED 系统的神经网络训练与测试的计算-能耗关联

From Computation to Consumption: Exploring the Compute-Energy Link for Training and Testing Neural Networks for SED Systems

University of Lorraine(洛林大学);CNRS(法国国家科学研究中心);Inria(法国国家信息与自动化研究所);Loria(洛林计算机科学与应用实验室)

Constance Douwes, Romain Serizel

AI总结:

本文针对声音事件检测系统中的神经网络,测量不同规模架构训练与测试的能耗,并揭示其与 FLOPs、参数量及 GPU/内存利用率的关系。

摘要:

机器学习模型,尤其是神经网络的大规模使用,已引发人们对其环境影响的严重担忧。事实上,在过去几年中,我们看到训练和部署这些系统所涉及的计算成本急剧增长。因此,了解其能源需求至关重要,以便将其更好地纳入模型评估;迄今为止,模型评估主要集中在性能上。本文以音频标注任务为例,研究了作为声音事件检测系统关键组成部分的若干神经网络架构。我们测量了从小型到大型架构在训练和测试阶段的能耗,并建立了能耗、浮点运算次数、参数数量以及 GPU/内存利用率之间的复杂关系。

英文摘要:

The massive use of machine learning models, particularly neural networks, has raised serious concerns about their environmental impact. Indeed, over the last few years we have seen an explosion in the computing costs associated with training and deploying these systems. It is, therefore, crucial to understand their energy requirements in order to better integrate them into the evaluation of models, which has so far focused mainly on performance. In this paper, we study several neural network architectures that are key components of sound event detection systems, using an audio tagging task as an example. We measure the energy consumption for training and testing small to large architectures and establish complex relationships between the energy consumption, the number of floating-point operations, the number of parameters, and the GPU/memory utilization.

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。

↑