The current Python example only returns chunk indexes and probabilities, which makes it hard to test the VAD model end-to-end on an actual audio file.
It would be great to have a simple code snippet that takes an audio file as input and returns voice activity timestamps — something like:
speech_timestamps = get_speech_timestamps(wav, model...)
Silero VAD has great quick start examples:
https://github.com/snakers4/silero-vad
The current Python example only returns chunk indexes and probabilities, which makes it hard to test the VAD model end-to-end on an actual audio file.
It would be great to have a simple code snippet that takes an audio file as input and returns voice activity timestamps — something like:
Silero VAD has great quick start examples:
https://github.com/snakers4/silero-vad