Sentence summarization system with Bi-encoder-decoder LSTM model

We publish the source code of the paper Deletion-based sentence compression using Bi-enc-dec LSTM.

@inproceedings{viet:2017:PACLING,
  author    = {Dac{-}Viet Lai and
               Nguyen Truong Son and
               Nguyen Le Minh},
  title     = {Deletion-Based Sentence Compression Using Bi-enc-dec {LSTM}},
  booktitle = {Computational Linguistics - 15th International Conference of the Pacific
               Association for Computational Linguistics, {PACLING} 2017, Yangon,
               Myanmar, August 16-18, 2017, Revised Selected Papers},
  pages     = {249--260},
  year      = {2017}
}

We built a web-based application and API from this model at our own server for English and Vietnamese. Please feel free to use for non-commercial purpose.

Dependencies

Python = 3.5

Tensorflow = 0.12.1

NLTK = 3.2.5

Gensim = 3.4.0

Stanford Postagger = 3.6

Installation

Clone the repository

git clone [email protected]:nguyenlab/SentSum.git

Install the dependencies

chmod +x install_dep.sh
./install_dep.sh

Preparing the data

Our system use CoNLL data format. We provide the 10.000 original-compressed pairs dataset provided by Katja Filippova in the data directory.

For vietnamese dataset, we can not publish it here. If you want to use it, contact us at [email protected].

If you want to use your own data, please convert yours into CoNLL format and config it in endata.py file.

Training

python run.py -train -i run1 

If you find any trouble, please raise an issue or contact us at [email protected] (main developer) or [email protected].