AudioRepInceptionNeXt: A lightweight single-stream architecture for efficient audio recognition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lau, Kin Wai, Rehman, Yasar Abbas Ur, Po, Lai-Man
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910417174921216
author Lau, Kin Wai
Rehman, Yasar Abbas Ur
Po, Lai-Man
author_facet Lau, Kin Wai
Rehman, Yasar Abbas Ur
Po, Lai-Man
contents Recent research has successfully adapted vision-based convolutional neural network (CNN) architectures for audio recognition tasks using Mel-Spectrograms. However, these CNNs have high computational costs and memory requirements, limiting their deployment on low-end edge devices. Motivated by the success of efficient vision models like InceptionNeXt and ConvNeXt, we propose AudioRepInceptionNeXt, a single-stream architecture. Its basic building block breaks down the parallel multi-branch depth-wise convolutions with descending scales of k x k kernels into a cascade of two multi-branch depth-wise convolutions. The first multi-branch consists of parallel multi-scale 1 x k depth-wise convolutional layers followed by a similar multi-branch employing parallel multi-scale k x 1 depth-wise convolutional layers. This reduces computational and memory footprint while separating time and frequency processing of Mel-Spectrograms. The large kernels capture global frequencies and long activities, while small kernels get local frequencies and short activities. We also reparameterize the multi-branch design during inference to further boost speed without losing accuracy. Experiments show that AudioRepInceptionNeXt reduces parameters and computations by 50%+ and improves inference speed 1.28x over state-of-the-art CNNs like the Slow-Fast while maintaining comparable accuracy. It also learns robustly across a variety of audio recognition tasks. Codes are available at https://github.com/StevenLauHKHK/AudioRepInceptionNeXt.
format Preprint
id arxiv_https___arxiv_org_abs_2404_13551
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AudioRepInceptionNeXt: A lightweight single-stream architecture for efficient audio recognition
Lau, Kin Wai
Rehman, Yasar Abbas Ur
Po, Lai-Man
Sound
Audio and Speech Processing
Recent research has successfully adapted vision-based convolutional neural network (CNN) architectures for audio recognition tasks using Mel-Spectrograms. However, these CNNs have high computational costs and memory requirements, limiting their deployment on low-end edge devices. Motivated by the success of efficient vision models like InceptionNeXt and ConvNeXt, we propose AudioRepInceptionNeXt, a single-stream architecture. Its basic building block breaks down the parallel multi-branch depth-wise convolutions with descending scales of k x k kernels into a cascade of two multi-branch depth-wise convolutions. The first multi-branch consists of parallel multi-scale 1 x k depth-wise convolutional layers followed by a similar multi-branch employing parallel multi-scale k x 1 depth-wise convolutional layers. This reduces computational and memory footprint while separating time and frequency processing of Mel-Spectrograms. The large kernels capture global frequencies and long activities, while small kernels get local frequencies and short activities. We also reparameterize the multi-branch design during inference to further boost speed without losing accuracy. Experiments show that AudioRepInceptionNeXt reduces parameters and computations by 50%+ and improves inference speed 1.28x over state-of-the-art CNNs like the Slow-Fast while maintaining comparable accuracy. It also learns robustly across a variety of audio recognition tasks. Codes are available at https://github.com/StevenLauHKHK/AudioRepInceptionNeXt.
title AudioRepInceptionNeXt: A lightweight single-stream architecture for efficient audio recognition
topic Sound
Audio and Speech Processing
url https://arxiv.org/abs/2404.13551