CoLA Dataset

07/29/2020 ∙ 0

DOWNLOAD CoLA

wget https://data.deepai.org/cola_public_1.1.zip
CoLA (Corpus of Linguistic Acceptability) consists of ~11k sentences from 23 unique linguistics publications. It is annotated for grammatical acceptability by their original authors. The public version included herein contains 9,594 sentences belonging to training and development sets, and excludes the 1063 sentences belonging to a held out test set.