Generates formal equivalents of informal input sentences. The model has been finetuned from a bart-base model using BLEU, BLEURT, and Style classifier rewards.
GitHub Repo