Training aborted due to the presence of an empty audio .
Created by: praveenjune17
Audio which has no word spoken in it can disrupt the training process. So before training ensure that the empty audios are removed. I used the below piece of code to identify the empty audios..
import soundfile as sf data, samplerate = sf.read('any_file.wav')
if sum(data) == 0: remove the audio from the training sample