Skip to content

Latest commit

 

History

History
45 lines (36 loc) · 1.36 KB

README.md

File metadata and controls

45 lines (36 loc) · 1.36 KB
Raw

Interactive Language Learning by Question Answering

Code for EMNLP 2019 paper "Interactive Language Learning by Question Answering".

To install dependencies

sudo apt update
conda create -p ~/venvs/qait python=3.6
source activate ~/venvs/qait
pip install --upgrade pip
pip install numpy==1.16.4
pip install https://github.com/Microsoft/TextWorld/archive/rebased-interactive-qa.zip
pip install -U spacy
python -m spacy download en
pip install tqdm h5py visdom pyyaml
conda install pytorch torchvision cudatoolkit=9.2 -c pytorch

Test Set

Download the test set from https://aka.ms/qait-testset. Unzip it.

Pretrained Word Embeddings

Before first time running it, download fasttext crawl-300d-2M.vec.zip from HERE, unzip, and run embedding2h5.py for fast embedding loading in the future.

To Train

python train.py ./

Citation

Please use the following bibtex entry:

@article{yuan2019qait,
  title={Interactive Language Learning by Question Answering},
  author={Yuan, Xingdi and C\^ot\'{e}, Marc-Alexandre and Fu, Jie and Lin, Zhouhan and Pal, Christopher and Bengio, Yoshua and Trischler, Adam},
  booktitle={EMNLP},
  year={2019}
}

License

MIT