News for hackers. Stuff that matters.

A comment on Audioscrape: Building in Rust When Everyone Said I Shouldn't. Parent comment

Re:Audioscrape: Building in Rust When Everyone Said I Shouldn't by lukaesch on Monday October 07, 2024 @12:35PM (#41767792)
whisper + pyannote + custom LLM pipeline to reduce errors, extract and attach speaker labels
[ Parent | on HN ]
Re:Audioscrape: Building in Rust When Everyone Said I Shouldn't by BetterWhisper on Monday October 07, 2024 @04:09PM (#41770437)
Are you running whisper on that same $7 Server?
[ Parent | on HN ]
Re:Audioscrape: Building in Rust When Everyone Said I Shouldn't by miki123211 on Monday October 07, 2024 @02:52PM (#41769558)
How do you interface with the Python ML stack from Rust?
[ Parent | on HN ]
Re:Audioscrape: Building in Rust When Everyone Said I Shouldn't by lukaesch on Monday October 07, 2024 @04:07PM (#41770405)
Whisper and Pyannote via Replicate to spin up a container for transcription and diarization. It returns word-level timestamps and speaker turns.

The LLM pipeline is built into the Rust code and makes calls to the OpenAI API.

[ Parent | on HN ]