Comment by elektor

Using the large model, it works really well, even in low volume settings/speakers mumbling. Some of my transcripts are pharma related and Whisper stumbles on the drug names, but I’m pretty understanding of that.