To run the model, you would need dependencies (e.g., vocab extracted from CoNLL corpus). For details, please refer to our [repo](https://github.com/utahnlp/structured_tuning_srl).