Open devanshkhandekar opened 1 year ago
I successfully detected synthetic Polish utterances using the provided pre-trained model.
However, please note that we based on the tiny.en
version of Whisper, which was trained only using English speech. Moreover, our DeepFake training dataset is not multi-lingual, so the results on other languages might be, therefore, worse.
Can this detection technique be used for multilingual synthetic audio detection ?