This is a project I recently came across from Prof Prathosh A. P. from IISc Bangalore.
Demos rendered with this system:
The TTS (text to speech) model was trained (fine tuned) on the professors voice input.
Paper and code is open source. It does only non-vedic chanting (out of choice).
Thought this may be interesting to some, hence sharing.
For anything further the professor can be contacted on the email on his site.
regards,
-Vardhan