Skip to content

Jessytan/Low-complexity-ASC

Repository files navigation

Low-complexity-ASC

This repository contains the introductions to the datasets and code used in our paper, titled "Low-Complexity Acoustic Scene Classification Using Parallel Attention-Convolution Network" (as shown in the section of Citation).

datasets

We conduct our experiments on the TAU Urban Acoustic Scene 2022 Mobile development dataset (TAU22) which consists of audio clips acquired by mobile devices in urban environments. You can download from here.

enviroment

The required library files are placed in requirements.txt
Our environment: RTX3090 + cuda11.3 + torch1.11

code

reuse

You need to create a directory named reuse to save training and validation data, and in the estimate_devices_freq.ipynb will used for spectrum modulation.

main

student.ipynb is the main code for training and validating the datasets.
Teacher models are same with the submission to the DCASE2023, and some are pretrained and provided in teacher_models.

Citation

Please cite our paper if you find the work in our paper are useful for your research.
[1] Y. Li, J. Tan, G. Chen, J. Li, Y. Si, and Q. He, "Low-Complexity Acoustic Scene Classification Using Parallel Attention-Convolution Network," in Proc. of Interspeech, 2024.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published