Usage

1. Add repository folder to PYTHONPATH: export PYTHONPATH=/this/repository/path.

2. Change the audio_folder item to yor audio dataset folder in config.yaml file.

3. At first time, preprocessing the audio data(wav file) to feature(melspectrogram, or stft) in npy format, save it in the ./data folder.

python3 preprocessing.py

4. Then start to training.

python3 main.py