Thanks a lot for maintaining this repo. It was really helpful for my research. I wanted to point out that the paper APPDIA: A Discourse-aware Transformer-based Style Transfer Model for Offensive Social Media Conversations, COLING 2022, fits better under the category Dataset or Supervised (Parallel Data). We release a dataset that contains 2K parallel pairs of offensive and style-transferred non-offensive texts. Our style transfer model is trained on this data in a supervised fashion. Our data can be found here: https://github.com/sabithsn/APPDIA-Discourse-Style-Transfer.
fuzhenxin
commented
2 years ago
Hi,
Thanks a lot for maintaining this repo. It was really helpful for my research. I wanted to point out that the paper APPDIA: A Discourse-aware Transformer-based Style Transfer Model for Offensive Social Media Conversations, COLING 2022, fits better under the category Dataset or Supervised (Parallel Data). We release a dataset that contains 2K parallel pairs of offensive and style-transferred non-offensive texts. Our style transfer model is trained on this data in a supervised fashion. Our data can be found here: https://github.com/sabithsn/APPDIA-Discourse-Style-Transfer.