Sovits icon indicating copy to clipboard operation
Sovits copied to clipboard

Published 20 hours ago • verified publisher icon Francis-Komizu

Sample Dataset

Open hopto-dot opened this issue 2 years ago • 1 comments

Your dataset should be like: └───wavs ├───dev │ ├───LJ001-0001.wav │ ├───... │ └───LJ050-0278.wav └───train ├───LJ002-0332.wav ├───... └───LJ047-0007.wav

Could you give me a link to see your dataset (AccelSpeech-Full.zip), or be more specific about what the difference between dev and train is. Also, why does train start off at 2 and 332 and end at 47 and 7?

Thanks

hopto-dot avatar Dec 05 '22 21:12 hopto-dot

As I understand it's an example from LJspeech dataset. I am also interested what's the minimum amount of wav files for train and dev?

Your dataset should be like: └───wavs ├───dev │ ├───LJ001-0001.wav │ ├───... │ └───LJ050-0278.wav └───train ├───LJ002-0332.wav ├───... └───LJ047-0007.wav

Could you give me a link to see your dataset (AccelSpeech-Full.zip), or be more specific about what the difference between dev and train is. Also, why does train start off at 2 and 332 and end at 47 and 7?

Thanks

stalevna avatar Dec 07 '22 09:12 stalevna