NTREX -- News Test References for MT Evaluation from English into a total of 128 target languages. See original GitHub repo for full details.
Example of loading:
dataset = load_dataset("davidstap/NTREX", "rus_Cyrl", trust_remote_code=True)
The following languages are available:
Language Code
Language Name
afr_Latn
Afrikaans
amh_Ethi
Amharic
arb_Arab
Arabic
aze_Latn
Azerbaijani
bak_Cyrl
Bashkir
bel_Cyrl