cardiffnlp/twitter-roberta-base-offensive

This is a roBERTa-base model trained on ~58M tweets and finetuned for offensive language identification with the TweetEval benchmark . The model was trained on 58M tweets . It is based on the findings of EMNLP 2020 . The data was downloaded from the official repository of Twitter-roberta-base . The code is used to train the model and test the accuracy of tweets with labels and labels . The results are based on tweets sent by users with their username and link placeholders . For example: "Not-offensive 0.9073.1" and "Good night" with "Good Night" with a "not-offensive" tag . The,