[Tacotron2] Other voices?

[turian]

I'm using Tacotron2 from torchhub.

Is there a simple way to change the speaker voice?

Is your feature request related to a problem? Please describe.

Want to work with Tacotron2 mel output from multiple speakers.

Describe the solution you'd like

There's a simple way to invoke the pre-trained model that switches the speaker voice. Since it was trained on librespeech, I assume there are several speakers in there.

Describe alternatives you've considered

Fine-tuning tacotron2 on new text-mel corpora.