Voicera is a smol text-to-speech program built for about 3 months by me, Lwasinam Dilli :). Voicera is open source, with trained model weights available at https://github.com/Lwasinam/voicera. It's not a SOTA model, just a smol project. I wanted to see if i could build something like this. This whoel project was inspired by James Betker, Hence why I cloned his demo code :). So if you ever see this, thank you.
This page demonstrates some of the results of Voicera.
Following are several particularly good results generated by the model.
p </body></html>