Questions about finetuning

[josephwong14wkh]

Thank you for your amazing work! I have 3 questions about finetuning the model.

1. Can this model clones the style of the target speaker, or it can only clone the speaker identity of the target speaker. (Actually I want to clone both speaker identity and style of target speaker)
2. Can i finetune the model on new languages, with 24kHz audio? Since I know the pretrained model is trained on 16kHz audio, can the model generalize to 24kHz audio after finetuning? And how many hours of data would you suggest for tuning the model on new languages?

Thank you!