olaverse/diacbench download history

olaverse/diacbench is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 72 times (18 in the last 7 days), and 593 times in total. It ranks #195,091 among datasets by monthly downloads.

DiacBench (v2) An 11-language benchmark for diacritics, tone-mark and Arabic vowel-mark restoration, built by Olaverse Lab to evaluate diacnet, diactag and any other diacritic restoration system on equal footing. Each config contains 1,000 sentence pairs: a diacritic-stripped input an

Models trained on diacbench

3 models list it as training data.

Open olaverse/diacbench on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.