dataset_raw/wav_structure.txt

数据集准备

raw
├───speaker0
│   ├───xxx1-xxx1.wav
│   ├───...
│   └───Lxx-0xx8.wav
└───speaker1
    ├───xx2-0xxx2.wav
    ├───...
    └───xxx7-xxx007.wav

此外还需要编辑config.json

"n_speakers": 10

"spk":{
    "speaker0": 0,
    "speaker1": 1,
}