Dear Mehrdad,
thanks a lot for your contributions, please i have a question, i saw your notebook to use wav2vec in Audio Classification and i want to use the model in speaker identification, if i trained the model with high accuracy and i need to add a new person to be classified will i need to re-run the whole training again with that huge amount of data, is the is any way to add new person’s voice to be identified with a couple of seconds of his voice data?