1 paper
Vít Novotný, Kristýna Luger, Michal Štefánik +2
Although pre-trained named entity recognition (NER) models are highly accurate on modern corpora, they underperform on historical texts due to differences in language OCR errors. I…