Final Model

• The final model is implemented in FINAL_MODEL_deberta_wrs_categorical.py.

Testing the Model

• The Final_Model_on_DevSet.ipynb notebook is used to:

  • Load the saved trained model.
  • Test it on the development dataset (dev_semeval_parids-labels.csv).
  • Generate labels using the trained model.

• Generated labels for devset is in dev.txt and for the hidden test set is in test.txt.

Training Logs

• The logs from training our final model (DeBERTa with weighted random sampling and categorical columns) are stored in FINAL_MODEL.log.

Usage

To train and save the final model, run the FINAL_MODEL_deberta_wrs_categorical.py. For testing and label generation on development and test datasets, execute the Final_Model_on_DevSet.ipynb notebook.