How to use it with non-parallel data?

[BenRafatian]

I was wondering that if I could use 2 sets of non-parallel voice datasets, where each speaker utters completely different sentences, and there is no transcript of the data.

also, does this program use end-end conversion, or does it train a speech-to-text model and then recreates the speech in the converted voice from the text?

[hikaruhotta]

MaskCycleGAN-VC is for non-parallel voice conversion so your dataset would work!

The model does end-to-end conversion meaning that it converts mel-spectrograms to mel-spectrograms

[BenRafatian]

Thank you @hikaruhotta for your help.