docker build --tag laituan245/spanbert_entity_coref .
docker push laituan245/spanbert_entity_coref
docker run --gpus '"device=0"' --rm -v /shared:/shared laituan245/spanbert_entity_coref -edl_official input1.tab -edl_freebase input2.tab -l ltf_dir -o output.tab
To train a model, simply run the following command
python runner.py
By default, the code will use the spanbert_large config.