PolyNews is a multilingual dataset containing news titles in 77 languages and 19 scripts.
This dataset can be used for domain adaptation of language models, language modeling or text generation.
There are 77 languages available:
Code
Language
Script
#Articles (K)
amh_Ethi
Amharic
Ethiopic
0.551
arb_Arab
Modern Standard Arabic
Arabic
10.882
ayr_Latn
Central Aymara
Latin