Roberta Large Cola Krishna2020

🚀 RoBERTa-large Classifier

A RoBERTa-large classifier trained on the CoLA corpus for evaluating the fluency of machine-generated English sentences.

🚀 Quick Start

This is a RoBERTa-large classifier trained on the CoLA corpus Warstadt et al., 2019. The CoLA corpus contains sentences paired with grammatical acceptability judgments. The model can be used to evaluate the fluency of machine-generated English sentences, for example, for the evaluation of text style transfer.

The model was trained in the paper Krishna et al, 2020. Reformulating Unsupervised Style Transfer as Paraphrase Generation, and its original version is available at their project page. We converted this model from Fairseq to the Transformers format. All credit goes to the authors of the original paper.

📚 Documentation

Citation

If you found this model useful and refer to it, please cite the original work:

@inproceedings{style20,
author={Kalpesh Krishna and John Wieting and Mohit Iyyer},
Booktitle = {Empirical Methods in Natural Language Processing},
Year = "2020",
Title={Reformulating Unsupervised Style Transfer as Paraphrase Generation},
}