This is a pytorch implemention of Gated-attention Reader Model(Gated-Attention Readers for Text)
- Python 3.6
- Pytorch 0.1.12_2
You can find data and pretained word embeddings file though this link.
You need to create your own data file, put data and pretrained word embedding file in it.
train dailymail data from raw:
python main.py --data_dir 'path to your data file' --dataset dailymail --embed_dir word2vec_glove.txt