jasonppy / VoiceCraft

Zero-Shot Speech Editing and Text-to-Speech in the Wild
Other
7.64k stars 746 forks source link

Total duration of training dataset #155

Open changjinhan opened 2 months ago

changjinhan commented 2 months ago

@jasonppy Thanks for sharing the open-source and enhanced version. But I am also curious about the amount of dataset. Could you let me know how many hours of data was used for training by combining Gigaspeech and Librilight?

jasonppy commented 2 months ago

About 20k hours

changjinhan commented 2 months ago

Thank you for your quick response!