Hierarchical Attention Networks for Document Classification

This is a 🔥 PyTorch 🔥 implementation of the fantastic HAN Paper from CMU(https://www.aclweb.org/anthology/N16-1174.pdf) complete with Yelp Dataloaders and Minibatching Loading for Training

Architecture

The only modification is that I have swapped out GloVe for Elmo for improved performance 🚀

Dataset

I have uploaded a sample dataset from the Yelp 2013 Reviews Dataset. Download the complete dataset at Yelp Reviews Dataset

Demo

I have also implemented a Streamlit App to interact with the model!

Instructions

To Install Requirements

pip install -r requirements.txt

To Train a New Model

python train.py

To Run the Demo Streamlit App

streamlit run app.py

Improvements

✌️ Pull requests are welcome for any improvements/features ✌️