English language, and Silence placeholder?

[alex32]

Is there a way to specify the language (English) so that whisper doesn't need to autodetect?

For example:

result = model.transcribe("short-02.mp3", language='English')

It would be great if Whisper could replace silence (pauses, say longer than 2sec) in the transcribed file. Eg. using a special character as a placeholder (~)

Thanks