Code and data of paper "Medical Dialogue Response Generation with Pivotal Information Recalling".
The final version is coming soon.
Train: python pretrain_and_recall.py
Inference: python meddialog.py
model: ./cikm_model/supervised_summary.py, config: main_config.py, trainer: supervised_summary_trainer.py, dataset: summary_response_dataset.py
Pretrain: pretrain_and_recall.py, main()