There are 4 folders.

DDA_Augmented contains the script to run Diversity Driven Attention model with Augmented Debatepedia Training Data. Most of the codes are collected from the official Repository

DDA_Original contains the script to run Diversity Driven Attention model without Augmented Debatepedia Training Data. Most of the codes are collected from the official Repository

PGN_DP contains the script to run Pointer Generation Network in Debatepedia Dataset. Most of the codes are collected from this Repository and the official Repository

BERTSUM_QFAS contains the script to run BERTSUM in Debatepedia Dataset. Most of the codes are collected from the official Repository

Instructions to run the codes are given in the official repository of each model (DDA, PGN, BERTSUM).

For DDA_Augmented and DDA_Original, 'data_fold' folder contains 10 fold cross validations data. The 'data' folder contains the dataset with given train-test-valid split for 1 fold experiment for few-shot learning to compare with PGN and BERTSUM.

For PGN_DP pre-processed bin files are given here to run the model in Debatepedia dataset.

For BERTSUM_QFAS, pre-processed python files are given here to run the model in Debatepedia dataset.