The repository is a reproduction of ACL 2017 paper Get To The Point: Summarization with Pointer-Generator Networks.
It heavily replied on the repository here.
I use the code to process CNN/DailyMail dataset.
Regarding model, i reconstruct the code using Tensorflow. It support (1) Pointer-Generator and (2) Coverage Mechanism
please see model.py for more details
Training validation and decode can be done concurrently.
please see run.py for more details.
It support python3.
please see config.py for more details.
ROUGE score, standard metric. I use the code.