Speker-Gender-Recognition

This project aims to build and evaluate several models that can classify speakers by their gender by audio files in Portuguese.

Virtual enviroment

To run the notebooks, first you will need to enter the virtual enviroment by running the following command:

    source ./venv/bin/activate

Datasets

• CETUC
• Multilingual TEDx
• MLS: Multilingual LibriSpeech (Obs. The gender label in this one is classified by another classifier, there could be some mistakes)