training from audiobook

Kyubyong / tacotron

A TensorFlow Implementation of Tacotron: A Fully End-to-End Text-To-Speech Synthesis Model

Apache License 2.0

1.83k stars 436 forks source link

Open pragmascript opened 6 years ago

pragmascript commented 6 years ago

Hello thank you for releasing this.

Could you expand on what preprocessing steps you had to take to be able to train from Nick's audiobooks if any?

davidmcgettigan commented 6 years ago

Hi, also interested to know how this was done as I would like to do this myself. @pragmascript did you figure out how they did it?

pragmascript commented 6 years ago

@HotSwap No... I have no idea.