kadirnar

/

Orpheus-TTS-Starrail

like 1

Hello,

May I ask you did you use to pass the speaker reference from the dataset during the training?

As seen in the dataset:

So I could reference them in inference, since the voice cloning using reference audio is broken right now...