ggerganov

/

whisper.cpp

like 786

Hello sir,
Thanks for this great models in a different version. I am trying to compare every model ACC and the whisper medium is more accurate for now. When I tried large the model output was like:

[00:00:00.000 --> 00:00:05.980]  [MUSIC PLAYING]
[00:00:05.980 --> 00:00:07.180]  [NON-ENGLISH SPEECH]
[00:00:07.180 --> 00:00:35.100]  [NON-ENGLISH SPEECH]
[00:00:35.100 --> 00:00:45.980]  [NON-ENGLISH SPEECH]

What is the issues here?