This repository contains a simple Question Answering algorithm developed in Tensorflow. We'll use the bAbI tasks 1-20, created by facebook, for testing the capacity of our algorithm to understand and reason the text.
It uses only a single GRU layer but perform better than the Facebook LSTM Baseline and Keras babi_rnn
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.
In order to run the notebook you need to install tensorflow and tqdm (to take progress).
$ pip install tensorflow --upgrade
$ pip install tqdm
Next, install git if you don't have already it and run the following commands on your terminal
$ cd $HOME # or any other development directory you prefer
$ git clone
$ cd bAbI
To keep track of the progress and to visualize the histograms of the NN layers, use Tensorboard, running on your terminal
$ cd $HOME # or any other development directory you prefer
$ cd bAbI
$ tensorboard --logdir tf_logs
Please let me know if you need more information or if you want to contribute in any way