Luyu/bert-base-mdoc-bm25

A text reranker trained for BM25 retriever on MS MARCO document dataset . It is possible to work with other retrievers like but using aligned BM25 works the best . We used anserini toolkit's BM25 implementation and indexed with tuned parameters (k1=3.8, b=0.87) following this instruction . See our project repo page for more details on how to use the toolkit and how-to-use it . The results of the project were published on Dev.com at the end of the 21st edition of the BERT Rerankers Project.com/Rethink Training of BERT.com . We are happy to,