README.md

metadata

license: mit
language:
  - zh
  - en
tags:
  - mteb
model-index:
  - name: bge-reranker-large-1k
    results:
      - task:
          type: Reranking
        dataset:
          type: C-MTEB/CMedQAv1-reranking
          name: MTEB CMedQAv1
          config: default
          split: test
          revision: None
        metrics:
          - type: map
            value: 82.15039301310942
          - type: mrr
            value: 84.92349206349208
      - task:
          type: Reranking
        dataset:
          type: C-MTEB/CMedQAv2-reranking
          name: MTEB CMedQAv2
          config: default
          split: test
          revision: None
        metrics:
          - type: map
            value: 84.18695515652256
          - type: mrr
            value: 86.96876984126985
      - task:
          type: Reranking
        dataset:
          type: C-MTEB/Mmarco-reranking
          name: MTEB MMarcoReranking
          config: default
          split: dev
          revision: None
        metrics:
          - type: map
            value: 37.641770303833106
          - type: mrr
            value: 36.5765873015873
      - task:
          type: Reranking
        dataset:
          type: C-MTEB/T2Reranking
          name: MTEB T2Reranking
          config: default
          split: dev
          revision: None
        metrics:
          - type: map
            value: 67.47867763060952
          - type: mrr
            value: 77.609724774586

本模型是在 BAAI/bge-reranker-large 上对模型的 model.base_model.embeddings.position_embeddings.weight 依赖经验值进行修改，来扩大模型输入长度，并没有进行任何继续supervised fined tuning，作为新手示例。 同时亦附上在C-MTEB的分数作为对比。 感谢原作者的工作。