GAN for Text Summarization

Tensorflow re-implementation of Generative Adversarial Network for Abstractive Text Summarization.

Requirements

• Python3

• Tensorflow >= 1.4 (tested on Tensorflow 1.4.1)

• numpy

• tqdm

• sklearn

• rouge

• pyrouge

You can use the python package manager of your choice (pip/conda) to install the dependencies. The code is tested on Ubuntu 16.04 operating system.

Quickstart

• Dataset

  Please follow the instructions here for downloading and preprocessing the CNN/DailyMail dataset. After that, store data files train.bin, val.bin, test.bin and vocabulary file vocab into specified data directory, e.g., ./data/.

• Prepare negative samples for discriminator

  You can download the prepared data discriminator_train_data.npz for discriminator from dropbox and store into specified data directory, e.g., ./data/.

• Train the full model

  python3 main.py --mode=train --data_path=./data/train.bin --vocab_path=./data/vocab --log_root=./log --pretrain_dis_data_path=./data/discriminator_train_data.npz --restore_best_model=False

• Decode

  python3 main.py --mode=decode --data_path=./data/test.bin --vocab_path=./data/vocab --log_root=./log --single_pass=True

References

• [1] Generative Adversarial Network for Abstractive Text Summarization (AAAI-18)

• [2] https://github.com/abisee/pointer-generator

• [3] https://github.com/LantaoYu/SeqGAN