arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2211.04772cs.SDcs.LGeess.AS
收藏

通过Transformer到CNN知识蒸馏的高效大规模音频标记

Efficient Large-scale Audio Tagging via Transformer-to-CNN Knowledge Distillation

Johannes Kepler University Linz(林茨约翰内斯·开普勒大学)

Florian Schmid, Khaled Koutini, Gerhard Widmer

AI总结:

本文提出利用离线知识蒸馏从Transformer训练高效CNN(基于MobileNetV3),在AudioSet上实现0.483 mAP的新最先进性能,兼顾参数、计算效率与预测性能。

摘要:

音频频谱图Transformer模型在音频标记领域占据主导地位,超越了先前占主导地位的卷积神经网络(CNN)。其优越性基于能够扩展并利用诸如AudioSet等大规模数据集。然而,与CNN相比,Transformer在模型大小和计算需求方面要求较高。我们提出了一种基于离线知识蒸馏(KD)的高效CNN训练流程,该流程从高性能但复杂的Transformer中蒸馏知识。所提出的训练方案以及基于MobileNetV3的高效CNN设计,使得模型在参数和计算效率以及预测性能方面优于先前的解决方案。我们提供了不同复杂度级别的模型,从低复杂度模型扩展到在AudioSet上达到新的最先进性能0.483 mAP。源代码可在:https://github.com/fschmid56/EfficientAT 获取。

英文摘要:

Audio Spectrogram Transformer models rule the field of Audio Tagging, outrunning previously dominating Convolutional Neural Networks (CNNs). Their superiority is based on the ability to scale up and exploit large-scale datasets such as AudioSet. However, Transformers are demanding in terms of model size and computational requirements compared to CNNs. We propose a training procedure for efficient CNNs based on offline Knowledge Distillation (KD) from high-performing yet complex transformers. The proposed training schema and the efficient CNN design based on MobileNetV3 results in models outperforming previous solutions in terms of parameter and computational efficiency and prediction performance. We provide models of different complexity levels, scaling from low-complexity models up to a new state-of-the-art performance of .483 mAP on AudioSet. Source Code available at: https://github.com/fschmid56/EfficientAT

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。

↑