license

#1
by wjbmattingly - opened

Hello, I am really interested in using this dataset to finetune a whisper model. Does the licensing permit this?

Owner

It can be used for fine-tuning, but the model trained or fine-tuned can not be commercialised.
Credit and citation must be given to NCHLT isXhosaSpeech Corpus.
[https://repo.sadilar.org/handle/20.500.12185/279]

Beijuka changed discussion status to closed

Thanks for the quick reply! I'm a huge fan of the work you are doing! Absolutely! I will be sure to link to the dataset on HF as well. If you like, I can keep you posted?

Owner

Thank you very much. I will be honoured.

I have the tiny, small, and medium models ready for Xhosa. I'm also training these same sizes on Zulu and Afrikaans. Finally, I have also created a new dataset that merges these into 30 second segments to reduce padding for Whisper. What should I cite for the Afrikaans and Zulu datasets besides your own work?

Thanks again!

Sign up or log in to comment