Skip to content

An annotated paper version of the GraphIE paper by Qian et al.

License

Notifications You must be signed in to change notification settings

polynoman/The-Annotated-GraphIE

Repository files navigation

The Annotated GraphIE

This annotated paper has been created as part of the course New Developments in NN Architectures and Representation Learning for NLU Tasks.

The original paper by Qian et al. can be found on arXiv.

The code is taken from the repository of the papers authors, which is available on Github.

This annotated paper contains only the word-level adaptation and application on the CoNLL-2003 dataset (Task 1).

Both code and paper text have been modified. The code has been put together and simplified for better understanding. The paper text has been rearranged to provide an explanation flow. Some parts that are not relevant have been removed. The license for the code remains the same (GPLv3).

My annotations are green and bold.

Installation

In this example, Python 3.6 is used.

  1. Install dependencies
pip install git+git://github.com/zhijing-jin/efficiency.git tensorboardX==1.4.0 scipy nltk jupyter
  1. Install Pytorch 0.4.0, for example via conda or using pip and a wheel (https://pytorch.org/get-started/previous-versions/)
  2. Download stopwords, run in python:
import nltk
nltk.download('stopwords')
  1. Download data
./download_data.sh

About

An annotated paper version of the GraphIE paper by Qian et al.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published