this post was submitted on 20 Jun 2026
34 points (85.4% liked)
Opensource
6302 readers
17 users here now
A community for discussion about open source software! Ask questions, share knowledge, share news, or post interesting stuff related to it!
⠀
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
So the training data was licensed with CC0 and is made available?
This seems to be a wrapper over Kokoro which is an Apache-licenced TTS model and library which has been out for over a year. That model lists about 10% of its approx 1000-hour training data but also says that all of its training data was CC, synthetic (by way of closed source TTs), or otherwise permissively-licensed.
huggingface kokoro
Interesting, does using the output of a nonfree software make the model more free?
Odd choice but personally I dont care, we have these models everywhere, please also on Linux etc.
Still looking for a good multi-lingual TTS on Android (english + german)