A repository implementation of the "BertSRC: BERT-based Semantic Relationship Classification" paper.

• Python >= 3.8
• Packages:

pip install -r requirements.txt

Train BertSRC train dataset

python train.py

Test BertSRC test dataset. Train must be run before the test.

python test.py

• kfold.ipynb

  • The notebook is an example of kfold evaluation for experiments mentioned in the paper.

• chemprot

  • The directory contains a chemprot dataset and a notebook that shows the results of the chemprot experiment.