Skip to content

benchmark

Earnings22

Public speech dataset of earnings-call audio used to test transcription accuracy on accented, jargon-heavy speech.

Current clusters