SpeechT5 Text to Speech fine tuning runtime error

did you find a fix to this or a work around?