Improve transcription speed on small files

[bakermanbrian]

I’ve been testing transcription speed for Whisper Base on small files (1-2 second clips) and have found performance is 3-7x slower than for larger audio files. Are there any optimizations to make to Whisper if I expect to exclusively transcribe smaller audio files?