Hi,
Thanks again for providing the Marathi resources. We include a Marathi NER model using your dataset as part of Stanza.
I am wondering, in the paper https://arxiv.org/pdf/2211.11418.pdf, you mention building a Gujarati Bert model, presumably using a large collection of raw Gujarati text. I found the Oscar Common Crawl and Wikipedia collections of Gujarati text. Did you use anything beyond that, and if so, are you able to make that text public in any way?
Thanks!
Hi,
Thanks again for providing the Marathi resources. We include a Marathi NER model using your dataset as part of Stanza.
I am wondering, in the paper https://arxiv.org/pdf/2211.11418.pdf, you mention building a Gujarati Bert model, presumably using a large collection of raw Gujarati text. I found the Oscar Common Crawl and Wikipedia collections of Gujarati text. Did you use anything beyond that, and if so, are you able to make that text public in any way?
Thanks!