Whisper

Whisper is a versatile speech recognition model, trained on diverse audio data and capable of multilingual recognition, translation, and language identification.