Skip to content

A chunker trained on the CoNLL 2000 task for chunking. It uses a BiLSTM plus CRF architecture.

Notifications You must be signed in to change notification settings

naveenjafer/Chunking-Keras-CoNLL-2000-BiLSTM-CRF-

Repository files navigation

Chunking-Keras-CoNLL-2000-BiLSTM-CRF

This is an implementation for the Chunking task as listed under CoNLL 2000 Dataset.

Architecture

It is a BiLstm and CRF implementation. The architecture is based off the paper titled Bidirectional LSTM-CRF Models for Sequence Tagging

Motivation

Rohit's repo on Named Entity Extraction using movies dataset was a very good starting point for this implementation. Some of the code has been used as is in the implementation.

Requirements

Tested with Python >= 1.7.0 & Python <= 1.15.0
Keras 2.2.4
Note: The code is written for a CPU implementation.

Evaluation

The script creates auxilarry files during the run for the predicted tags of the input sentences. Although the model predicts all the tags, I have only implemented evaluation of Precision, Recall and F1Score for Noun Phrase chunks. You can easily extend this to also evaluate Verb Phrases and PPN etc.

Instructions to run

I have included the jupyter notebook file and the corresponding python3 vanilla file version of the same.

Jupyter notebook

jupyter notebook

python3 script

python3 BiLstm_+_crf_for_chunking.py

Future Work

Working on making this compatible with Tensorflow > 2.0 and corresponding Keras versions.

About

A chunker trained on the CoNLL 2000 task for chunking. It uses a BiLSTM plus CRF architecture.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published