BERT Reranker for MS-MARCO Document Ranking

Model description

A text reranker trained for BM25 retriever on MS MARCO document dataset.

Intended uses & limitations

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.

How to use

See our project repo page.

Eval results

MRR @10: 0.423 on Dev.

BibTeX entry and citation info

               title={Rethink Training of BERT Rerankers in Multi-Stage Retrieval Pipeline}, 
               author={Luyu Gao and Zhuyun Dai and Jamie Callan},
               booktitle={The 43rd European Conference On Information Retrieval (ECIR)},
Downloads last month
Hosted inference API
Text Classification
This model can be loaded on the Inference API on-demand.