CoNLL 2003 Tagger

This repo contains a tagger on CoNLL2003 dataset. The data contains three files, training set, testing set and validation set. The data contains POS tags, Chunk tags and Named Entity Tags. No pre-trained model and no pre-trained embedding have been used. All three tagging tasks have been done using Bi-LSTM models. The data can be found here.

Pre-requisites

Use requirements.txt file to install the required packages. You could also create a new environment using this file.

pip install -r requirements.txt

Usage

After downloading/cloning this repo, open your terminal. Change directory to the root of the repo and run the following to see the Jupyter Notebook.

jupyter notebook conll2003_analysis.ipynb

Files

The data files are in a folder called "data". All other files are in the root.

Version

Python 3.7.1

Author

Upasana Parashar