Whisper is a general-purpose speech transcription model. It is trained on a large dataset of diverse audio and is also a multi-task model that can perform multilingual speech transcription as well as speech translation and language identification.

This version uses the lasts whisper version available and add a new input to perform the transcription.