We can find the article here
We can find resources here
To train the model it is necessary to put the folder of the audio files in the same folder that train.py and to change the name of nom_du_dossier in the file outils_entrainement.py
And then we can train the model with :
python3 train.py
To test it is necessary to put the audio files in the folder examples and execute :
python3 test.py
If you want to calculate again the weigth change b to False in test.py