bash download_data.sh
https://drive.google.com/drive/folders/1jVyTfXx4qFusggaZlSxYGe96WrRUKqxU?usp=sharing
data
|- ccp
| |- config.json
| |- pytorch_model.bin
| |- sentencepece.bpe.model
| |- tokenizer.json
|- nq
| |- biencoder-nq-dev.json
| |- biencoder-nq-train.json
|- xorqa
| |- dev.jsonl
| |- en_wiki.tsv
| |- test.jsonl
| |- train.jsonl
|- mrtydi
| |- ar
| | |- collection
| | | |- docs.jsonl
| | |- pid2passage.tsv
| | |- qrels.dev.txt
| | |- qrels.test.txt
| | |- qrels.train.txt
| | |- qrels.txt
| | |- topic.dev.tsv
| | |- topic.test.tsv
| | |- topic.train.tsv
| | |- topic.tsv
| |- bn
...
pip install -r requirements.txt
bash run-ccp.sh
python eval/mrtydi.py --pred_file runs/CCP/mrtydi --set test
python eval/xorqa.py --pred_file runs/CCP/xorqa/dev.json