This repository contains Jupyter Notebooks with our team SSN_NLP's code for the OffensEval 2020 Task 12: Offense Target Identification in Social Media.
The original code for fine-tuning and pre-training BERT can be found at this repository and the datasets are available at the challenge's OffensEval page.
We achieved an F1 score of 0.61 over our best run that uses BERT (Bidirectional Encoder Representations from Transformers) for the purpose of training the OLI model