Joint Learning of Emotions in Music and Generalized Sounds

Note

If you use our work in your research, cite us:

Simonetta F., Certo F., Ntalampiras S. "Joint Learning of Emotions in Music and Generalized Sounds", AudioMostly 2024, Milan, Italy. DOI: https://doi.org/10.1145/3678299.3678324

Reproduce

Using pdm

This project was developed using pdm and intel MKL libraries. To setup the same identical environment, do as follows:

1. Install pdm (e.g. using pipx)
2. Enter the project directory and run pdm sync
3. Download IADS-E and PMEmo datasets ad unzip them
4. Download OpenSmile 3.0.1
5. Download and extract the datasets each in a different directory (IADS-E, IADS-2, PMEmo)
6. Modify music_sound_emotions/settings.py to match your paths:

• the path to OpenSmile root directory
• the paths to the datasets root directories

8. From the project root run:

• pdm features to extract features
• pdm experiment to reproduce our experiments
• pdm parse experiment*.log to parse the logs