Can Whisper do a line break or new line for a voice change?

[cyberbrix]

I was running through a documentary and despite the clear changes in voice, Whisper output everything into a single line for multiple voices, with no breaks.

Is there a good option or setting to help improve that?

[ryanheise]

There are a couple of approaches:

1. Use Pyannote: Transcription and diarization (speaker identification) #264
2. Use prompt magic: prompt vs prefix in DecodingOptions #117 (comment)