speechbrain

/

vad-crdnn-libriparty

like 23

Hi,I'm running this code:

audio_path = "end_framework_it's_also_fair_work_its.wav"
signal, fs = torchaudio.load(audio_path)
signal = signal.squeeze()
VAD = VAD.from_hparams(source="speechbrain/vad-crdnn-libriparty", savedir="pretrained_models/vad-crdnn-libriparty")
probs = VAD.get_speech_prob_file(audio_path)

Returns all probs to be 1.0, I'm reading it from file 16khz, what can be the problem?
Thanks !