Inroduction

This is the source code of our paper "A Two-student Learning Framework for Mixed Supervised Target Sound Detection" The paper can be found on https://arxiv.org/pdf/2204.02088.pdf

Data

For reference audio, you can directly download from our google drive https://drive.google.com/file/d/1PwTDj0yIStc_NKPn9RxgZGsTOZ7kGjfA/view?usp=sharing
For mixture audio, please download Audioset dataset, URBAN-SED dataset, and then using our scripts to generate data.

How to run it

You can extract the feature according our code, or directly using our extracted feature (we will release it soon). Afther that you can directly run "bash bash/*.sh" .The different experiments can be choose from bash folder.

To do list

(1) we will release the extracted feature and our paper as soon.
(2) Re-organize the code, make it easy to read.
(3) If you have any problem, please contact me without hesitation.