Basic Info

Author: Rujun Han*, I-Hung Hsu*, Mu Yang

Title: Codebase for CoNLL 2019 Paper: Deep Structured Neural Network for Event Temporal Relation Extraction

Data processinng. We have preprocessed MATRES(notice that the Matres dataset we use are their initial released version, hence, contains less data), TB-Dense and TCR raw data using internal NLP tools at the Information Sciences Institute. These .pickle files are saved in data fold.

Additional Note:

• If you are curious about the data preprocessing, we recommend you to see this script : https://github.com/rujunhan/TEDataProcessing/blob/master/processMATRES.py
• However, this is not been test in this repo, thus, the best case is to download data from the link we provide

Setup

1. Install Gurobi

conda config --add channels http://conda.anaconda.org/gurobi
conda install gurobi
grbgetkey $YOURGUROBIKEY$

2. Return to this repository and:

pip install -r requirements.txt

Or see the yml file we append

3. Download data.

Run code:

cd Code
python train_all.py