I’m new in Machine learning. I have task to live transcribe audio from microphone and also make word-level transcripts of this live transcription instead of utterance level.