Datasets:

---

muzaffercky

/

kurdish-web

like 1

Hi Muzaffer — really impressed by Kurdish-Web and Kurdish-Kurmanji-News,
especially the download numbers, that's rare for Kurdish NLP resources.

We've been building something in a similar space: KurdishCorpus-clean
(huggingface.co/datasets/kurdish-tech/KurdishCorpus-clean) — 2.7M
documents / 2.97B tokens across Kurmancî, Soranî and Zazakî, deduplicated
and quality-filtered.

Would be great if it fit into your Kurdish Collection alongside your
other resources — either way, thanks for the work you're putting into
this space.