JAX port of the Dia TTS model (1.6B params) from Nari Labs for inference on any machine.
pip install diajax==0.0.7
dia --text "Hey, I'm really sorry for getting back to you so late. (cough) But voice cloning is just super easy, it's barely an inconvenience at all. I will show you how." --audio "assets/example_prompt.mp3"
pip install diajax==0.0.7 dia --text "Hey, I'm really sorry for getting back to you so late. (cough) But voice cloning is just super easy, it's barely an inconvenience at all. I will show you how." --audio "assets/example_prompt.mp3"