The repository to generete the LitSet dataset for our EACL'24 paper.
LitSet uses entity disambiguation datasets for label interpretation training. The resulting model can be used for few-shot named entity recognition using verbalized label descriptions.
conda create -name litset_dataset python=3.10
conda activate litset_dataset
pip install -r requirements.txt
python create_dataset.py --download_data --output_format jsonl