I've been trying to replace the webrtcvad present in the coqui stt example present here with the silero vad. While using the silero vad with .wav files, the detection of speech samples happens perfectly. Once I try to implement the same in the coqui stt example, I hit bottlenecks whereby only part of my speech is transcribed.
I've tried playing with the utils.py file of the silero vad to replace the below line but it was not of much help.
is_speech = self.vad.is_speech(frame, self.sample_rate)
I tried multiple combinations but I am unable to reach proper transcribing, appreciate if someone could help out.