I am trying to start training a WaveRNN vocoder, but I am having all sorts of trouble with the notebook that extracts spectrograms using a TTS model. What is the difference if we use preprocess_data.py instead?
I am trying to start training a WaveRNN vocoder, but I am having all sorts of trouble with the notebook that extracts spectrograms using a TTS model. What is the difference if we use preprocess_data.py instead?