This is a fine-tuned version of BERT trained on the dataset of real and fake news used for indicating a reliability of the news article.
Additional Model Description
The intended use for this model is to output a percentage from 0 to 100 indicating the reliability of the news article. It was specifically developed for my project that combines fact checking on the internet with the prediction from this model. This model is basically a fine-tuned version of BERT, trained on a merged dataset consisting of two Kaggle datasets: one for real news and one for fake news.
