textattack/roberta-base-WNLI

The model was fine-tuned for sequence classification using TextAttack and the glue dataset loaded using the nlp library . The best score the model achieved on this task was 0.5633802816901409, as measured by the eval set accuracy, found after 0 epoch . The model had a learning rate of 5e-05 and a maximum sequence length of 256 . It was trained with a cross-entropy loss function, and the data was loaded with a dataset loaded with nlp . For more information, check out TextAttack on Github and read the latest article in this week's open-source version of this article by Andrew Keen and his colleagues at the Open Daybreak,