Stanford Question Answering Dataset

What is this repository for?

• This repository implements two model trained on the SQuAD dataset, one for question answering, and one for information retrieval.

Requirements

• Python 3
• Keras
• Tensorflow
• Numpy
• Pickle 3
• Pandas

Overview

• E2E.ipynb Contains the module for End to End information retrieval
• SQuAD_project.ipynb Contains the base SQuAD model
• IR_SQuAD.ipynb Contains the module for information retrieval
• compute_answers.py A script that takes as input a dataset containing questions and contexts and returns the predicted answers.

How do I run the model?

• ./python3 compute_answers.py PATH_TO_DATASET

Who do I talk to?

• Andrea Lavista - [email protected]
• Daniele Domenichelli - [email protected]
• Luca Rispoli - [email protected]