Code for paper "QAScore - An Unsupervised Unreferenced Metric for the Question Generation Evaluation" (see this link)
- python=3.8
- transformers
- pytorch
- tqdm
run the following command
python -u QAScore.py \
--paragraph_file example_files/paragraphs.txt \
--question_file example_files/questions.txt \
--answer_file example_files/answers.txt \
--device "cuda" \
--show_progress
and the result is -0.9693889460924203