---

Exploring Multi-Descriptor Disentangled Representation of Acoustic Instrument Notes

Francesco Ardan Dal Rí - Gregorio Andrea Giudici - Luca Turchet - Nicola Conci

Repository Structure

./multidescriptor-vae/
├── data/
├── docs/
├── models/
├── notebooks/
├── scripts/
├── src/
├── requirements.txt
└── README.md

Get Started

Clone it using git:

git clone https://github.com/gregogiudici/multidescriptor-vae.git
cd multidescriptor-vae

Create a venv

# Create venv
python -m venv venv
# Activate venv
venv\Scripts\activate # In Windows
source venv/bin/activate # In Linux/MacOS
# Install required packages
pip install - requirements.txt

Preprocessing the TinySOL Dataset

To preprocess the TinySOL dataset, modify the ORIGIN_DATASET_FOLDER and TARGET_DATASET_FOLDER variables, and run:

python dataset/tinysol_preprocessing.py

Then, create the dataframe with:

python dataset/create_dataframe.py

Training the Multi-Descriptor GMVAE

To train the Multi-Descriptor GMVAE, use the following command:

python src/train_multidescriptor.py