Text2NKG: Fine-Grained N-ary Relation Extraction for N-ary relational Knowledge Graph Construction [paper].
- Linux(SSH) + Python3.7.13 + Pytorch1.8.1 + Cuda11.1
pip install torch==1.8.1+cu111 torchvision==0.9.1+cu111 torchaudio==0.8.1 -f https://download.pytorch.org/whl/torch_stable.html
Install dependencies, The code is based on huggaface's transformers:
pip install -r requirement.txt
pip install --editable ./transformers
Then, we should install apex.
unzip apex.zip
cd apex
python setup.py install
cd ..
Our experiments are based on HyperRED dataset.
Download Pre-trained Language Models from Hugging Face:
mkdir -p bert_models/bert-base-uncased
wget -P bert_models/bert-base-uncased https://huggingface.co/bert-base-uncased/resolve/main/pytorch_model.bin
wget -P bert_models/bert-base-uncased https://huggingface.co/bert-base-uncased/resolve/main/vocab.txt
wget -P bert_models/bert-base-uncased https://huggingface.co/bert-base-uncased/resolve/main/config.json
mkdir -p bert_models/bert-large-uncased
wget -P bert_models/bert-large-uncased https://huggingface.co/bert-large-uncased/resolve/main/pytorch_model.bin
wget -P bert_models/bert-large-uncased https://huggingface.co/bert-large-uncased/resolve/main/vocab.txt
wget -P bert_models/bert-large-uncased https://huggingface.co/bert-large-uncased/resolve/main/config.json
wget -P bert_models/bert-large-uncased https://huggingface.co/bert-large-uncased/resolve/main/tokenizer.json
wget -P bert_models/bert-large-uncased https://huggingface.co/bert-large-uncased/resolve/main/tokenizer_config.json
Train Text2NKG n-ary RE Models:
python run_re.py
If you find this work is helpful for your research, please cite:
@misc{luo2023text2nkg,
title={Text2NKG: Fine-Grained N-ary Relation Extraction for N-ary relational Knowledge Graph Construction},
author={Haoran Luo and Haihong E and Yuhao Yang and Tianyu Yao and Yikai Guo and Zichen Tang and Wentai Zhang and Kaiyang Wan and Shiyao Peng and Meina Song and Wei Lin},
year={2023},
eprint={2310.05185},
archivePrefix={arXiv},
primaryClass={cs.AI}
}
For further questions, please contact: [email protected], or wechat: lhr1846205978.