File size: 521 Bytes
8c960e3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
174ec59
7924ca8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
from customs.make_custom_dataset import create_dataset

'''
How should the data_dir be created?
Place the necessary audio files in data_dir.
Transcription, tokenization, etc. of the audio files are done by the create_dataset function.

data_dir
├── bpe_69.json
├── utt1.wav
├── utt2.wav
├── utt3.wav
......
└── utt{n}.wav
'''

data_dir = "VALL-E-X_Dataset/data_dir"
create_dataset(data_dir, audio_path, filename, transcription, duration, sample_rate, asr_model, dataloader_process_only=True)