Support for long-form synthesis.

[computersrmyfriends]

The gradio app displays that
"MetaVoice-1B is a 1.2B parameter base model for TTS (text-to-speech). It has been built with the following priorities:

 **Support for long-form synthesis.

and on the main page:

Is there some way to do it now which is probably not documented?

[MethanJess]

@computersrmyfriends
Long-form support is still in the works, see #53

[vatsalaggarwal]

Is there some way to do it now which is probably not documented?

No, model is built in a way to handle this, but extra code is needed to make it work... we'll release it along with other upcoming features

[computersrmyfriends]

For now, is it possible to split text and generate it?

[vatsalaggarwal]

Yes, you can do that manually... but it will mean that the audio sounds inconsistent across sentences...

[computersrmyfriends]

Still waiting for your long form support. Please let me know if you have any updates

[lucapericlp]

Thanks for the enthusiasm & patience @computersrmyfriends, we'll post an update here once we have more to talk about 👍

[fivestones]

I'm excited for this feature too!