GitHub Typo Corpus

AbstractThe lack of large-scale datasets has been a major hindrance to the development of NLP tasks such as spelling correction and grammatical error correction (GEC). As a complementary new resource …...