XL-Sum

Abstractive Summarization for African Languages

summarization text-generation sequence-to-sequence 6k-7k pairs per language CC-BY-NC-SA 4.0

Description

Large-scale abstractive summarization dataset from BBC News. High-quality article-summary pairs for many African languages.

Quick Start

Python
from datasets import load_dataset

dataset = load_dataset("GEM/xlsum", "igbo", split="train")
print(dataset['text'][:2], dataset['target'][:2])

Project Ideas

mT5 Summarization

Fine-tune mT5 to summarize Igbo news articles

Intermediate

ROUGE Evaluation

Evaluate ROUGE scores on Amharic text

Beginner

Ethical Considerations

  • BBC data represents specific editorial viewpoint