Can this be tweaked as a TTS only model with very high intonations?

[tensorjackal]

Hey! Amazing work. I have a usecase which requires expressive TTS. Can I fine-tune this on HQ 44kHz TTS data and make it work as TTS only but having intonations as much as in singing?