Skip to content

This code is for paper "OTExtSum: Extractive Text Summarisation with Optimal Transport", Findings of NAACL 2022

License

Notifications You must be signed in to change notification settings

peggypytang/OTExtSum

Repository files navigation

OTExtSum

This code is for paper OTExtSum: Extractive Text Summarisation with Optimal Transport, Findings of NAACL 2022

To test the performance with BIP optimisation strategy using BERT representation:

python OTExtSum_BIP.py --save_path=decoded_bip_cnndm3 --dataset_str cnn_dailymail --dataset_doc_field article --max_output_length 3 --tkner bert --device cuda

To test the performance with BS optimisation strategy using BERT representation:

python OTExtSum_BS.py --save_path=decoded_bs_pubmed6 --dataset_str pubmed --dataset_doc_field article --tkner bert --max_output_length 6 --device cuda

About

This code is for paper "OTExtSum: Extractive Text Summarisation with Optimal Transport", Findings of NAACL 2022

Resources

License

Stars

Watchers

Forks

Languages