crimson storm
Just to confirm, do Datasets refer to collections of audio files used to train or fine-tune a Voice in Voice Pro Cloning?
If so, am I able to upload multiple files AFTER I've already generated a voice to train it better?
And if this is possible? does it train itself with the new recordings or do I need to "tell" the cloner?
tulip troutBOT
Hey @crimson storm,
Thank you for reaching out to Cartesia Support!
That is correct! Datasets are collections of audio files used to fine-tune a model. You can add files to an existing dataset; however, a Pro Voice Clone (PVC) already trained will not update automatically. To incorporate new files, you need to retrain and create a new PVC using the updated dataset.
crimson storm
Hi Ege, thanks for getting back to me, how do I retrain an existing voice? do i need to make a new voice in PVC?