This method includes a Python script and files of a RoBERTa-base model fine-tuned on our Cybersecurity NER dataset. The method requires as input a list of sentences that will be fed to the model. The method's output is a list of predictions, i.e. the labels assigned to each token of each of the sentences fed to the model. INPUT= sentences: string[] OUTPUT= predictions: list(list(dict(string, string))) The transformers model RoBERTa was pretrained in a self-supervised manner using a sizable corpus of English data. This indicates that it was pretrained using only the raw texts and an algorithmic procedure to create inputs and labels from those texts. No human labeling was done at all, which allows it to use a large amount of publicly available data.