CoolFace
Datasetpublic

codemurt/uyghur_ner_dataset

Uyghur NER dataset Description This dataset is in WikiAnn format. The dataset is assembled from named entities parsed from Wikipedia, Wiktionary and Dbpedia. For some words, new case forms have been created using Apertium-uig. Some locations have been translated using the Google Translate API. The dataset is divided into two parts: train and extra. Train has full sentences, extra has only named entities. Tags: O (0), B-PER (1), I-PER (2), B-ORG (3), I-ORG (4)… See the full description on the dataset page: https://huggingface.co/datasets/codemurt/uyghur_ner_dataset.

sourceHugging Facemitupdated 3y agoView on Hugging Face
3likes20downloads
Dataset Card

Uyghur NER dataset

Description

This dataset is in WikiAnn format. The dataset is assembled from named entities parsed from Wikipedia, Wiktionary and Dbpedia. For some words, new case forms have been created using Apertium-uig. Some locations have been translated using the Google Translate API.

The dataset is divided into two parts: train and extra. Train has full sentences, extra has only named entities.

Tags: O (0), B-PER (1), I-PER (2), B-ORG (3), I-ORG (4), B-LOC (5), I-LOC (6)

Data example

{
	'tokens': ['قاراماي', 'شەھىرى', '«مەملىكەت', 'بويىچە', 'مىللەتل…'],
 	'ner_tags': [5, 0, 0, 0, 0],
 	'langs': ['ug', 'ug', 'ug', 'ug', 'ug'],
 	'spans': ['LOC: قاراماي']
}

Usage with datasets library

py
from datasets import load_dataset

dataset = load_dataset("codemurt/uyghur_ner_dataset")