Request for open-sourcing data-processing code
#1
by
DotIN13
- opened
The Chinese wikipedia indeed is a fuss to process, taking into account the messy multilingual template tags, and deeply nested HTML tags, etc. I'm wondering if you would kindly open-source the processing code for convenient customization of future wiki dumps, to facilitate the training of more Chinese foundational models on up-to-date high-quality wiki data. Many thanks.