I was transcribing some audios that I have (I’m sorry but I’m not able to share them) with whisper-tiny (@openai ?) and for some of them the output looked crazy: it got stuck in some words and kept repeating again and again the same. Some examples:
1- “the state of the state of the state of the state of the…”
2- “it’s it’s it’s it’s it’s …”
I’m quite confident that I did the right thing (I was batch transcribing and everything else worked fine) and I just wanted to flag this out.
I’ve tried with bigger models (in my case, whisper-medium) and it works fine with it.
I hope that the information I provide is good enough and/or that you are aware of the bug already.
Thanks for all the models, they work really nice 