SumArabic

Published: 29 June 2022| Version 1 | DOI: 10.17632/7kr75c9h24.1
Contributor:
Mohammad Bani Almarjeh

Description

SumArabic is an Arabic abstractive text summarization dataset. The uploaded files are: - sumarabic-1.0-index.jsonl.xz: a file that contains the Common Crawl records - A script to download the data The data are from the following two Arabic news websites: - emaratalyoum.com - almamlakatv.com The data are splitted into training, testing, validation, and out-of-domain sets. The number of examples in each split is as follows: Training: 75,817 Validation: 4,121 Testing: 4,174 Out-of-domain: 652 Total: 84,764

Files

Categories

Arabic Language, Text Processing

License