vespa-engine/colbert-medium

The model is based on ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT . The official baseline BM25 ranking model MRR@10 is 0.16 on eval and 0.167 on dev question set . The model weights have been tuned by training using the triples.train.small.tar.gz from MSMARCO-Passage-Ranking.ai . To use this model with vespa.ai for MS Marco Passage Ranking, see  referiorMS Marco Ranking using Vespa .ai sample app. The model has been exported to ONNX for serving in Vespa's support for running models. For,