|
|
+> Transcribing inference audio for computing metric uses [Whisper](https://github.com/openai/whisper#setup), which is automatically installed. Whisper in turn requires the command-line tool [`ffmpeg`](https://ffmpeg.org/) to be installed on your system, which is available from most package managers.
|