ttslearn
v0.2.2 (2022-01-04)
Japanese continues below (Japanese followers)
English: This book is written in Japanese and primary focuses on Japanese TTS. Some of the functionality (eg, neural network implementations) in this codebase can be used for other languages. However, we didn't prepare any guide or code for non-Japanese TTS systems. We may extend the codebase for other languages in the future but cannot guarantee if we would work on it.
pip install ttslearn
pip install ttslearn . It can be used not only as sample code for books, but also as a general-purpose speech synthesis library.ttslearn library is used, the repository includes recipes for speech synthesis using the JSUT corpus and JVS corpus.For more detailed documentation, see https://r9y9.github.io/ttslearn/.
The source code is licensed by MIT. Can be used for both commercial and non-commercial purposes. For more information, see the LICENSE file.
The release page of this repository distributes trained models that are learned using the JSUT corpus and JVS corpus. Those trained models are only available for "non-commercial purposes". When using a trained model, please also check the terms of use for each corpus.
Furthermore, the author shall not be liable for any claims, damages or other obligations due to the use of the learned model.
As an appendix, we have compiled the specifications for full context labels for Japanese speech synthesis. For more information, see docs/appendix.pdf.
If you have any questions about the book's contents or source code, please contact us via GitHub issue and we will respond as much as possible.
The errata for this book is summarized at the link below.
Errata of this book
If you find any errors such as typographical errors not listed in the errata, please contact us via GitHub issue.
