---
license: cc-by-4.0
language:
- ur
pretty_name: Sangraha Verified Urdu
source_datasets:
- ai4bharat/sangraha
multilinguality:
- monolingual
---

# Sangraha Verified Urdu

This is the `verified/urd` portion of the
[Sangraha](https://huggingface.co/datasets/ai4bharat/sangraha/) dataset. The
data is tokenized to sentences using
[UrduHack's](https://github.com/urduhack/urduhack) `sentence_tokenizer` with
one sentence per row. Original metadata is not retained.