Hierarchical Attention Network for Document Classification NAACL 2016

This repository contains Keras implementation of Hierarchical Attention Networks for Document Classification(NAACL 2016)

Requirements:-

• Keras>=2.1.x
• Tensorflow>=1.4.x
• Glove embeddings

Run train.ipynb to get final sentiment labels.

The paper implemented is available here.

This code uses data from Kaggle which is available at this link.