Voice data <= 10 mins can also be used to train a good VC model! - View it on GitHub