Datasets:
seperation of text sources
#4
by
eddy-mm
- opened
Hi.
thanks for your great effort for creating dataset for middle east languages.
As i understood from your descriptions, you crawled it from multiple sources.
is there a way that i can separate it?
Hey Eddy, thank you.
We originally had that idea to make it reproducible. Therefore we built https://huggingface.co/datasets/SLPL/naab-raw, but some of the links are not there yet or they are inaccessible. Anyways I would like to refer you to our pre-print: https://arxiv.org/abs/2208.13486. There you can find link to the open sources we used for building naab.
Unfortuanletly there isn't anyway you can separate the dataset by source.