Does BEIR support tuning config k and b of BM25 ?

[Yuan0320]

Hi @thakur-nandan, thanks for the amazing work! I would like to know if BEIR supports tuning config k and b of BM25? In the paper, anserini BM25 was implemented with default Lucene parameters (k=0.9 and b=0.4). I want to run evaluate_bm25.py or evaluate_anserini_bm25.py of beir with different k and b. But I can't find where to tune them.

Thanks.

[lintool]

You can find Anserini regressions for BEIR here: https://github.com/castorini/anserini#regressions-for-beir-v100 You can reuse the tuning script implemented here: https://github.com/castorini/anserini/blob/master/docs/experiments-msmarco-passage.md#bm25-tuning

[Jia-py]

You can find Anserini regressions for BEIR here: https://github.com/castorini/anserini#regressions-for-beir-v100 You can reuse the tuning script implemented here: https://github.com/castorini/anserini/blob/master/docs/experiments-msmarco-passage.md#bm25-tuning

Does BEIR support originally setting the k and b for BM25, or should I import the BM25 from Anserini to set these parameters? I find the scripts in the two links are hard to follow. Are there any other relevant tutorials or documentation?

Thanks