Extract Speakers Entities

[nhtkid]

Dear Community,

I am using OpenAI Whisper API to extract Podcast into transcriptions.

The result is a blob of text with no punctuation or paragraph.

Are there any parameters I could use to get a better result?

Ideally, for a typical Podcast conversation, I like the output to be like dialog.

For example, list each entity in front of each sentence.

Guest 1: ......
Guest 2: ......
Guest 1: ......

Any suggestion is welcome.
Thanks all.

Happy New Year 2023

[glangford]

This has already been discussed in
#104
#340
#571
among other places. Search for "diarization" in the Discussions. See also the recently announced project called Podalize.
https://github.com/mave5/podalize

[Majdoddin]

Hi @nhtkid @glangford ,
www.lexicaps.com seamlessly adds diarization to Whispers transcription. No 3rd party packages.
Feel free to contact me if you need more features.
Announcement: #1537
Repo: https://github.com/Majdoddin/lexicaps