Are the training datasets publicly available?

[Majdoddin]

Thank you for amazing and open source Whisper 👍

As I understand it from the paper, you collected the proper training data from internet, and then did some preprocessing on it (re-sampling, ...). Is this training data publicly available?
I need it in the process of saprsifying the whisper's model using sparseeml. Of course I could use standard datasets that are used for evaluation of whisper, but whisper's robustness is based on using a diverse training dataset.

[JaejinCho]

I had the same question, too although the collected data does not seem open to the public.

[phineas-pta]

unfortunately no, even the training code aren't public (they only provide fine tune code)