This helps transcription accuracy and tagging.

Specify if you know how many speakers are in the audio. Improves diarization accuracy.

Separates and labels different speakers in the transcription. Requires GPU and takes longer.