Spaces:

openai

/

whisper

like 1.4k

Running on T4

Hello,

I am aiming to fine-tune whisper in the Jordanian dialect, I was wondering if I would be still using the Arabic tokenizer?

Also, do you have any advice on how to collect voice recordings for training data?

I hope to take this into the next step in transcribing mixed-language speech ( English and Arabic)

Would appreciate any advice,

Best