Replies: 1 comment 1 reply
-
audio_lengths is tensor(2) probably means that you have 2 channels in your audio. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hey!
I am working with this notebook for diarizing the audio files. This notebook seems to work perfectly with the given "an4_diarize_test.wav" test file but when I try it on my own audio files then model gives runtime error as shown below. I wonder why it is not processing the audio properly and giving the empty tensor. My audio files are also in .wav format, therefore, there is not any inconsistency as far as audio file format is concerned.
I also went through the source code of audio_to_label.py and I found that audio_signal tensor has 0 elements. Moreover, the audio_lengths is tensor(2) which means it is only processing two samples of audio which does not make sense. Any help would be much appreciated!
Beta Was this translation helpful? Give feedback.
All reactions