Sentence Similarity
sentence-transformers
PyTorch
English
bert
feature-extraction
mteb
custom_code
Eval Results
6 papers
willathum's picture
Update README
e33bf52
metadata
license: apache-2.0
pipeline_tag: sentence-similarity
inference: false
tags:
  - sentence-transformers
  - feature-extraction
  - sentence-similarity
  - mteb
language: en
datasets:
  - s2orc
  - flax-sentence-embeddings/stackexchange_title_body_jsonl
  - flax-sentence-embeddings/stackexchange_titlebody_best_voted_answer_jsonl
  - flax-sentence-embeddings/stackexchange_title_best_voted_answer_jsonl
  - >-
    flax-sentence-embeddings/stackexchange_titlebody_best_and_down_voted_answer_jsonl
  - sentence-transformers/reddit-title-body
  - msmarco
  - gooaq
  - yahoo_answers_topics
  - code_search_net
  - search_qa
  - eli5
  - snli
  - multi_nli
  - wikihow
  - natural_questions
  - trivia_qa
  - embedding-data/sentence-compression
  - embedding-data/flickr30k-captions
  - embedding-data/altlex
  - embedding-data/simple-wiki
  - embedding-data/QQP
  - embedding-data/SPECTER
  - embedding-data/PAQ_pairs
  - embedding-data/WikiAnswers
  - sentence-transformers/embedding-training-data
model-index:
  - name: lodestone-base-4096-v1
    results:
      - task:
          type: Classification
        dataset:
          type: mteb/amazon_counterfactual
          name: MTEB AmazonCounterfactualClassification (en)
          config: en
          split: test
          revision: e8379541af4e31359cca9fbcf4b00f2671dba205
        metrics:
          - type: accuracy
            value: 69.7313432835821
          - type: ap
            value: 31.618259511417733
          - type: f1
            value: 63.30313825394228
      - task:
          type: Classification
        dataset:
          type: mteb/amazon_polarity
          name: MTEB AmazonPolarityClassification
          config: default
          split: test
          revision: e2d317d38cd51312af73b3d32a06d1a08b442046
        metrics:
          - type: accuracy
            value: 86.89837499999999
          - type: ap
            value: 82.39500885672128
          - type: f1
            value: 86.87317947399657
      - task:
          type: Classification
        dataset:
          type: mteb/amazon_reviews_multi
          name: MTEB AmazonReviewsClassification (en)
          config: en
          split: test
          revision: 1399c76144fd37290681b995c656ef9b2e06e26d
        metrics:
          - type: accuracy
            value: 44.05
          - type: f1
            value: 42.67624383248947
      - task:
          type: Retrieval
        dataset:
          type: arguana
          name: MTEB ArguAna
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 26.173999999999996
          - type: map_at_10
            value: 40.976
          - type: map_at_100
            value: 42.067
          - type: map_at_1000
            value: 42.075
          - type: map_at_3
            value: 35.917
          - type: map_at_5
            value: 38.656
          - type: mrr_at_1
            value: 26.814
          - type: mrr_at_10
            value: 41.252
          - type: mrr_at_100
            value: 42.337
          - type: mrr_at_1000
            value: 42.345
          - type: mrr_at_3
            value: 36.226
          - type: mrr_at_5
            value: 38.914
          - type: ndcg_at_1
            value: 26.173999999999996
          - type: ndcg_at_10
            value: 49.819
          - type: ndcg_at_100
            value: 54.403999999999996
          - type: ndcg_at_1000
            value: 54.59
          - type: ndcg_at_3
            value: 39.231
          - type: ndcg_at_5
            value: 44.189
          - type: precision_at_1
            value: 26.173999999999996
          - type: precision_at_10
            value: 7.838000000000001
          - type: precision_at_100
            value: 0.9820000000000001
          - type: precision_at_1000
            value: 0.1
          - type: precision_at_3
            value: 16.287
          - type: precision_at_5
            value: 12.191
          - type: recall_at_1
            value: 26.173999999999996
          - type: recall_at_10
            value: 78.378
          - type: recall_at_100
            value: 98.222
          - type: recall_at_1000
            value: 99.644
          - type: recall_at_3
            value: 48.862
          - type: recall_at_5
            value: 60.953
      - task:
          type: Clustering
        dataset:
          type: mteb/arxiv-clustering-p2p
          name: MTEB ArxivClusteringP2P
          config: default
          split: test
          revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d
        metrics:
          - type: v_measure
            value: 42.31689035788179
      - task:
          type: Clustering
        dataset:
          type: mteb/arxiv-clustering-s2s
          name: MTEB ArxivClusteringS2S
          config: default
          split: test
          revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53
        metrics:
          - type: v_measure
            value: 31.280245136660984
      - task:
          type: Reranking
        dataset:
          type: mteb/askubuntudupquestions-reranking
          name: MTEB AskUbuntuDupQuestions
          config: default
          split: test
          revision: 2000358ca161889fa9c082cb41daa8dcfb161a54
        metrics:
          - type: map
            value: 58.79109720839415
          - type: mrr
            value: 71.79615705931495
      - task:
          type: STS
        dataset:
          type: mteb/biosses-sts
          name: MTEB BIOSSES
          config: default
          split: test
          revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
        metrics:
          - type: cos_sim_pearson
            value: 76.44918756608115
          - type: cos_sim_spearman
            value: 70.86607256286257
          - type: euclidean_pearson
            value: 74.12154678100815
          - type: euclidean_spearman
            value: 70.86607256286257
          - type: manhattan_pearson
            value: 74.0078626964417
          - type: manhattan_spearman
            value: 70.68353828321327
      - task:
          type: Classification
        dataset:
          type: mteb/banking77
          name: MTEB Banking77Classification
          config: default
          split: test
          revision: 0fd18e25b25c072e09e0d92ab615fda904d66300
        metrics:
          - type: accuracy
            value: 75.40584415584415
          - type: f1
            value: 74.29514617572676
      - task:
          type: Clustering
        dataset:
          type: mteb/biorxiv-clustering-p2p
          name: MTEB BiorxivClusteringP2P
          config: default
          split: test
          revision: 65b79d1d13f80053f67aca9498d9402c2d9f1f40
        metrics:
          - type: v_measure
            value: 37.41860080664014
      - task:
          type: Clustering
        dataset:
          type: mteb/biorxiv-clustering-s2s
          name: MTEB BiorxivClusteringS2S
          config: default
          split: test
          revision: 258694dd0231531bc1fd9de6ceb52a0853c6d908
        metrics:
          - type: v_measure
            value: 29.319217023090705
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackAndroidRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 26.595000000000002
          - type: map_at_10
            value: 36.556
          - type: map_at_100
            value: 37.984
          - type: map_at_1000
            value: 38.134
          - type: map_at_3
            value: 33.417
          - type: map_at_5
            value: 35.160000000000004
          - type: mrr_at_1
            value: 32.761
          - type: mrr_at_10
            value: 41.799
          - type: mrr_at_100
            value: 42.526
          - type: mrr_at_1000
            value: 42.582
          - type: mrr_at_3
            value: 39.39
          - type: mrr_at_5
            value: 40.727000000000004
          - type: ndcg_at_1
            value: 32.761
          - type: ndcg_at_10
            value: 42.549
          - type: ndcg_at_100
            value: 47.915
          - type: ndcg_at_1000
            value: 50.475
          - type: ndcg_at_3
            value: 37.93
          - type: ndcg_at_5
            value: 39.939
          - type: precision_at_1
            value: 32.761
          - type: precision_at_10
            value: 8.312
          - type: precision_at_100
            value: 1.403
          - type: precision_at_1000
            value: 0.197
          - type: precision_at_3
            value: 18.741
          - type: precision_at_5
            value: 13.447999999999999
          - type: recall_at_1
            value: 26.595000000000002
          - type: recall_at_10
            value: 54.332
          - type: recall_at_100
            value: 76.936
          - type: recall_at_1000
            value: 93.914
          - type: recall_at_3
            value: 40.666000000000004
          - type: recall_at_5
            value: 46.513
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackEnglishRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 22.528000000000002
          - type: map_at_10
            value: 30.751
          - type: map_at_100
            value: 31.855
          - type: map_at_1000
            value: 31.972
          - type: map_at_3
            value: 28.465
          - type: map_at_5
            value: 29.738
          - type: mrr_at_1
            value: 28.662
          - type: mrr_at_10
            value: 35.912
          - type: mrr_at_100
            value: 36.726
          - type: mrr_at_1000
            value: 36.777
          - type: mrr_at_3
            value: 34.013
          - type: mrr_at_5
            value: 35.156
          - type: ndcg_at_1
            value: 28.662
          - type: ndcg_at_10
            value: 35.452
          - type: ndcg_at_100
            value: 40.1
          - type: ndcg_at_1000
            value: 42.323
          - type: ndcg_at_3
            value: 32.112
          - type: ndcg_at_5
            value: 33.638
          - type: precision_at_1
            value: 28.662
          - type: precision_at_10
            value: 6.688
          - type: precision_at_100
            value: 1.13
          - type: precision_at_1000
            value: 0.16
          - type: precision_at_3
            value: 15.562999999999999
          - type: precision_at_5
            value: 11.019
          - type: recall_at_1
            value: 22.528000000000002
          - type: recall_at_10
            value: 43.748
          - type: recall_at_100
            value: 64.235
          - type: recall_at_1000
            value: 78.609
          - type: recall_at_3
            value: 33.937
          - type: recall_at_5
            value: 38.234
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackGamingRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 33.117999999999995
          - type: map_at_10
            value: 44.339
          - type: map_at_100
            value: 45.367000000000004
          - type: map_at_1000
            value: 45.437
          - type: map_at_3
            value: 41.195
          - type: map_at_5
            value: 42.922
          - type: mrr_at_1
            value: 38.37
          - type: mrr_at_10
            value: 47.786
          - type: mrr_at_100
            value: 48.522
          - type: mrr_at_1000
            value: 48.567
          - type: mrr_at_3
            value: 45.371
          - type: mrr_at_5
            value: 46.857
          - type: ndcg_at_1
            value: 38.37
          - type: ndcg_at_10
            value: 50.019999999999996
          - type: ndcg_at_100
            value: 54.36299999999999
          - type: ndcg_at_1000
            value: 55.897
          - type: ndcg_at_3
            value: 44.733000000000004
          - type: ndcg_at_5
            value: 47.292
          - type: precision_at_1
            value: 38.37
          - type: precision_at_10
            value: 8.288
          - type: precision_at_100
            value: 1.139
          - type: precision_at_1000
            value: 0.132
          - type: precision_at_3
            value: 20.293
          - type: precision_at_5
            value: 14.107
          - type: recall_at_1
            value: 33.117999999999995
          - type: recall_at_10
            value: 63.451
          - type: recall_at_100
            value: 82.767
          - type: recall_at_1000
            value: 93.786
          - type: recall_at_3
            value: 48.964999999999996
          - type: recall_at_5
            value: 55.358
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackGisRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 16.028000000000002
          - type: map_at_10
            value: 23.186999999999998
          - type: map_at_100
            value: 24.236
          - type: map_at_1000
            value: 24.337
          - type: map_at_3
            value: 20.816000000000003
          - type: map_at_5
            value: 22.311
          - type: mrr_at_1
            value: 17.514
          - type: mrr_at_10
            value: 24.84
          - type: mrr_at_100
            value: 25.838
          - type: mrr_at_1000
            value: 25.924999999999997
          - type: mrr_at_3
            value: 22.542
          - type: mrr_at_5
            value: 24.04
          - type: ndcg_at_1
            value: 17.514
          - type: ndcg_at_10
            value: 27.391
          - type: ndcg_at_100
            value: 32.684999999999995
          - type: ndcg_at_1000
            value: 35.367
          - type: ndcg_at_3
            value: 22.820999999999998
          - type: ndcg_at_5
            value: 25.380999999999997
          - type: precision_at_1
            value: 17.514
          - type: precision_at_10
            value: 4.463
          - type: precision_at_100
            value: 0.745
          - type: precision_at_1000
            value: 0.101
          - type: precision_at_3
            value: 10.019
          - type: precision_at_5
            value: 7.457999999999999
          - type: recall_at_1
            value: 16.028000000000002
          - type: recall_at_10
            value: 38.81
          - type: recall_at_100
            value: 63.295
          - type: recall_at_1000
            value: 83.762
          - type: recall_at_3
            value: 26.604
          - type: recall_at_5
            value: 32.727000000000004
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackMathematicaRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 11.962
          - type: map_at_10
            value: 17.218
          - type: map_at_100
            value: 18.321
          - type: map_at_1000
            value: 18.455
          - type: map_at_3
            value: 15.287999999999998
          - type: map_at_5
            value: 16.417
          - type: mrr_at_1
            value: 14.677000000000001
          - type: mrr_at_10
            value: 20.381
          - type: mrr_at_100
            value: 21.471999999999998
          - type: mrr_at_1000
            value: 21.566
          - type: mrr_at_3
            value: 18.448999999999998
          - type: mrr_at_5
            value: 19.587
          - type: ndcg_at_1
            value: 14.677000000000001
          - type: ndcg_at_10
            value: 20.86
          - type: ndcg_at_100
            value: 26.519
          - type: ndcg_at_1000
            value: 30.020000000000003
          - type: ndcg_at_3
            value: 17.208000000000002
          - type: ndcg_at_5
            value: 19.037000000000003
          - type: precision_at_1
            value: 14.677000000000001
          - type: precision_at_10
            value: 3.856
          - type: precision_at_100
            value: 0.7889999999999999
          - type: precision_at_1000
            value: 0.124
          - type: precision_at_3
            value: 8.043
          - type: precision_at_5
            value: 6.069999999999999
          - type: recall_at_1
            value: 11.962
          - type: recall_at_10
            value: 28.994999999999997
          - type: recall_at_100
            value: 54.071999999999996
          - type: recall_at_1000
            value: 79.309
          - type: recall_at_3
            value: 19.134999999999998
          - type: recall_at_5
            value: 23.727999999999998
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackPhysicsRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 22.764
          - type: map_at_10
            value: 31.744
          - type: map_at_100
            value: 33.037
          - type: map_at_1000
            value: 33.156
          - type: map_at_3
            value: 29.015
          - type: map_at_5
            value: 30.434
          - type: mrr_at_1
            value: 28.296
          - type: mrr_at_10
            value: 37.03
          - type: mrr_at_100
            value: 37.902
          - type: mrr_at_1000
            value: 37.966
          - type: mrr_at_3
            value: 34.568
          - type: mrr_at_5
            value: 35.786
          - type: ndcg_at_1
            value: 28.296
          - type: ndcg_at_10
            value: 37.289
          - type: ndcg_at_100
            value: 42.787
          - type: ndcg_at_1000
            value: 45.382
          - type: ndcg_at_3
            value: 32.598
          - type: ndcg_at_5
            value: 34.521
          - type: precision_at_1
            value: 28.296
          - type: precision_at_10
            value: 6.901
          - type: precision_at_100
            value: 1.135
          - type: precision_at_1000
            value: 0.152
          - type: precision_at_3
            value: 15.367
          - type: precision_at_5
            value: 11.03
          - type: recall_at_1
            value: 22.764
          - type: recall_at_10
            value: 48.807
          - type: recall_at_100
            value: 71.859
          - type: recall_at_1000
            value: 89.606
          - type: recall_at_3
            value: 35.594
          - type: recall_at_5
            value: 40.541
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackProgrammersRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 19.742
          - type: map_at_10
            value: 27.741
          - type: map_at_100
            value: 29.323
          - type: map_at_1000
            value: 29.438
          - type: map_at_3
            value: 25.217
          - type: map_at_5
            value: 26.583000000000002
          - type: mrr_at_1
            value: 24.657999999999998
          - type: mrr_at_10
            value: 32.407000000000004
          - type: mrr_at_100
            value: 33.631
          - type: mrr_at_1000
            value: 33.686
          - type: mrr_at_3
            value: 30.194
          - type: mrr_at_5
            value: 31.444
          - type: ndcg_at_1
            value: 24.657999999999998
          - type: ndcg_at_10
            value: 32.614
          - type: ndcg_at_100
            value: 39.61
          - type: ndcg_at_1000
            value: 42.114000000000004
          - type: ndcg_at_3
            value: 28.516000000000002
          - type: ndcg_at_5
            value: 30.274
          - type: precision_at_1
            value: 24.657999999999998
          - type: precision_at_10
            value: 6.176
          - type: precision_at_100
            value: 1.1400000000000001
          - type: precision_at_1000
            value: 0.155
          - type: precision_at_3
            value: 13.927
          - type: precision_at_5
            value: 9.954
          - type: recall_at_1
            value: 19.742
          - type: recall_at_10
            value: 42.427
          - type: recall_at_100
            value: 72.687
          - type: recall_at_1000
            value: 89.89
          - type: recall_at_3
            value: 30.781
          - type: recall_at_5
            value: 35.606
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 19.72608333333333
          - type: map_at_10
            value: 27.165333333333336
          - type: map_at_100
            value: 28.292499999999997
          - type: map_at_1000
            value: 28.416333333333327
          - type: map_at_3
            value: 24.783833333333334
          - type: map_at_5
            value: 26.101750000000003
          - type: mrr_at_1
            value: 23.721500000000002
          - type: mrr_at_10
            value: 30.853333333333328
          - type: mrr_at_100
            value: 31.741750000000003
          - type: mrr_at_1000
            value: 31.812999999999995
          - type: mrr_at_3
            value: 28.732249999999997
          - type: mrr_at_5
            value: 29.945166666666665
          - type: ndcg_at_1
            value: 23.721500000000002
          - type: ndcg_at_10
            value: 31.74883333333333
          - type: ndcg_at_100
            value: 36.883583333333334
          - type: ndcg_at_1000
            value: 39.6145
          - type: ndcg_at_3
            value: 27.639583333333334
          - type: ndcg_at_5
            value: 29.543666666666667
          - type: precision_at_1
            value: 23.721500000000002
          - type: precision_at_10
            value: 5.709083333333333
          - type: precision_at_100
            value: 0.9859166666666666
          - type: precision_at_1000
            value: 0.1413333333333333
          - type: precision_at_3
            value: 12.85683333333333
          - type: precision_at_5
            value: 9.258166666666668
          - type: recall_at_1
            value: 19.72608333333333
          - type: recall_at_10
            value: 41.73583333333334
          - type: recall_at_100
            value: 64.66566666666668
          - type: recall_at_1000
            value: 84.09833333333336
          - type: recall_at_3
            value: 30.223083333333328
          - type: recall_at_5
            value: 35.153083333333335
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackStatsRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 17.582
          - type: map_at_10
            value: 22.803
          - type: map_at_100
            value: 23.503
          - type: map_at_1000
            value: 23.599999999999998
          - type: map_at_3
            value: 21.375
          - type: map_at_5
            value: 22.052
          - type: mrr_at_1
            value: 20.399
          - type: mrr_at_10
            value: 25.369999999999997
          - type: mrr_at_100
            value: 26.016000000000002
          - type: mrr_at_1000
            value: 26.090999999999998
          - type: mrr_at_3
            value: 23.952
          - type: mrr_at_5
            value: 24.619
          - type: ndcg_at_1
            value: 20.399
          - type: ndcg_at_10
            value: 25.964
          - type: ndcg_at_100
            value: 29.607
          - type: ndcg_at_1000
            value: 32.349
          - type: ndcg_at_3
            value: 23.177
          - type: ndcg_at_5
            value: 24.276
          - type: precision_at_1
            value: 20.399
          - type: precision_at_10
            value: 4.018
          - type: precision_at_100
            value: 0.629
          - type: precision_at_1000
            value: 0.093
          - type: precision_at_3
            value: 9.969
          - type: precision_at_5
            value: 6.748
          - type: recall_at_1
            value: 17.582
          - type: recall_at_10
            value: 33.35
          - type: recall_at_100
            value: 50.219
          - type: recall_at_1000
            value: 71.06099999999999
          - type: recall_at_3
            value: 25.619999999999997
          - type: recall_at_5
            value: 28.291
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackTexRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 11.071
          - type: map_at_10
            value: 16.201999999999998
          - type: map_at_100
            value: 17.112
          - type: map_at_1000
            value: 17.238
          - type: map_at_3
            value: 14.508
          - type: map_at_5
            value: 15.440999999999999
          - type: mrr_at_1
            value: 13.833
          - type: mrr_at_10
            value: 19.235
          - type: mrr_at_100
            value: 20.108999999999998
          - type: mrr_at_1000
            value: 20.196
          - type: mrr_at_3
            value: 17.515
          - type: mrr_at_5
            value: 18.505
          - type: ndcg_at_1
            value: 13.833
          - type: ndcg_at_10
            value: 19.643
          - type: ndcg_at_100
            value: 24.298000000000002
          - type: ndcg_at_1000
            value: 27.614
          - type: ndcg_at_3
            value: 16.528000000000002
          - type: ndcg_at_5
            value: 17.991
          - type: precision_at_1
            value: 13.833
          - type: precision_at_10
            value: 3.6990000000000003
          - type: precision_at_100
            value: 0.713
          - type: precision_at_1000
            value: 0.116
          - type: precision_at_3
            value: 7.9030000000000005
          - type: precision_at_5
            value: 5.891
          - type: recall_at_1
            value: 11.071
          - type: recall_at_10
            value: 27.019
          - type: recall_at_100
            value: 48.404
          - type: recall_at_1000
            value: 72.641
          - type: recall_at_3
            value: 18.336
          - type: recall_at_5
            value: 21.991
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackUnixRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 18.573
          - type: map_at_10
            value: 25.008999999999997
          - type: map_at_100
            value: 26.015
          - type: map_at_1000
            value: 26.137
          - type: map_at_3
            value: 22.798
          - type: map_at_5
            value: 24.092
          - type: mrr_at_1
            value: 22.108
          - type: mrr_at_10
            value: 28.646
          - type: mrr_at_100
            value: 29.477999999999998
          - type: mrr_at_1000
            value: 29.57
          - type: mrr_at_3
            value: 26.415
          - type: mrr_at_5
            value: 27.693
          - type: ndcg_at_1
            value: 22.108
          - type: ndcg_at_10
            value: 29.42
          - type: ndcg_at_100
            value: 34.385
          - type: ndcg_at_1000
            value: 37.572
          - type: ndcg_at_3
            value: 25.274
          - type: ndcg_at_5
            value: 27.315
          - type: precision_at_1
            value: 22.108
          - type: precision_at_10
            value: 5.093
          - type: precision_at_100
            value: 0.859
          - type: precision_at_1000
            value: 0.124
          - type: precision_at_3
            value: 11.474
          - type: precision_at_5
            value: 8.321000000000002
          - type: recall_at_1
            value: 18.573
          - type: recall_at_10
            value: 39.433
          - type: recall_at_100
            value: 61.597
          - type: recall_at_1000
            value: 84.69
          - type: recall_at_3
            value: 27.849
          - type: recall_at_5
            value: 33.202999999999996
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackWebmastersRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 22.807
          - type: map_at_10
            value: 30.014000000000003
          - type: map_at_100
            value: 31.422
          - type: map_at_1000
            value: 31.652
          - type: map_at_3
            value: 27.447
          - type: map_at_5
            value: 28.711
          - type: mrr_at_1
            value: 27.668
          - type: mrr_at_10
            value: 34.489
          - type: mrr_at_100
            value: 35.453
          - type: mrr_at_1000
            value: 35.526
          - type: mrr_at_3
            value: 32.477000000000004
          - type: mrr_at_5
            value: 33.603
          - type: ndcg_at_1
            value: 27.668
          - type: ndcg_at_10
            value: 34.983
          - type: ndcg_at_100
            value: 40.535
          - type: ndcg_at_1000
            value: 43.747
          - type: ndcg_at_3
            value: 31.026999999999997
          - type: ndcg_at_5
            value: 32.608
          - type: precision_at_1
            value: 27.668
          - type: precision_at_10
            value: 6.837999999999999
          - type: precision_at_100
            value: 1.411
          - type: precision_at_1000
            value: 0.23600000000000002
          - type: precision_at_3
            value: 14.295
          - type: precision_at_5
            value: 10.435
          - type: recall_at_1
            value: 22.807
          - type: recall_at_10
            value: 43.545
          - type: recall_at_100
            value: 69.39800000000001
          - type: recall_at_1000
            value: 90.706
          - type: recall_at_3
            value: 32.183
          - type: recall_at_5
            value: 36.563
      - task:
          type: Retrieval
        dataset:
          type: BeIR/cqadupstack
          name: MTEB CQADupstackWordpressRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 13.943
          - type: map_at_10
            value: 20.419999999999998
          - type: map_at_100
            value: 21.335
          - type: map_at_1000
            value: 21.44
          - type: map_at_3
            value: 17.865000000000002
          - type: map_at_5
            value: 19.36
          - type: mrr_at_1
            value: 15.712000000000002
          - type: mrr_at_10
            value: 22.345000000000002
          - type: mrr_at_100
            value: 23.227999999999998
          - type: mrr_at_1000
            value: 23.304
          - type: mrr_at_3
            value: 19.901
          - type: mrr_at_5
            value: 21.325
          - type: ndcg_at_1
            value: 15.712000000000002
          - type: ndcg_at_10
            value: 24.801000000000002
          - type: ndcg_at_100
            value: 29.799
          - type: ndcg_at_1000
            value: 32.513999999999996
          - type: ndcg_at_3
            value: 19.750999999999998
          - type: ndcg_at_5
            value: 22.252
          - type: precision_at_1
            value: 15.712000000000002
          - type: precision_at_10
            value: 4.1770000000000005
          - type: precision_at_100
            value: 0.738
          - type: precision_at_1000
            value: 0.106
          - type: precision_at_3
            value: 8.688
          - type: precision_at_5
            value: 6.617000000000001
          - type: recall_at_1
            value: 13.943
          - type: recall_at_10
            value: 36.913000000000004
          - type: recall_at_100
            value: 60.519
          - type: recall_at_1000
            value: 81.206
          - type: recall_at_3
            value: 23.006999999999998
          - type: recall_at_5
            value: 29.082
      - task:
          type: Retrieval
        dataset:
          type: climate-fever
          name: MTEB ClimateFEVER
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 9.468
          - type: map_at_10
            value: 16.029
          - type: map_at_100
            value: 17.693
          - type: map_at_1000
            value: 17.886
          - type: map_at_3
            value: 13.15
          - type: map_at_5
            value: 14.568
          - type: mrr_at_1
            value: 21.173000000000002
          - type: mrr_at_10
            value: 31.028
          - type: mrr_at_100
            value: 32.061
          - type: mrr_at_1000
            value: 32.119
          - type: mrr_at_3
            value: 27.534999999999997
          - type: mrr_at_5
            value: 29.431
          - type: ndcg_at_1
            value: 21.173000000000002
          - type: ndcg_at_10
            value: 23.224
          - type: ndcg_at_100
            value: 30.225
          - type: ndcg_at_1000
            value: 33.961000000000006
          - type: ndcg_at_3
            value: 18.174
          - type: ndcg_at_5
            value: 19.897000000000002
          - type: precision_at_1
            value: 21.173000000000002
          - type: precision_at_10
            value: 7.4719999999999995
          - type: precision_at_100
            value: 1.5010000000000001
          - type: precision_at_1000
            value: 0.219
          - type: precision_at_3
            value: 13.312
          - type: precision_at_5
            value: 10.619
          - type: recall_at_1
            value: 9.468
          - type: recall_at_10
            value: 28.823
          - type: recall_at_100
            value: 53.26499999999999
          - type: recall_at_1000
            value: 74.536
          - type: recall_at_3
            value: 16.672
          - type: recall_at_5
            value: 21.302
      - task:
          type: Retrieval
        dataset:
          type: dbpedia-entity
          name: MTEB DBPedia
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 6.343
          - type: map_at_10
            value: 12.717
          - type: map_at_100
            value: 16.48
          - type: map_at_1000
            value: 17.381
          - type: map_at_3
            value: 9.568999999999999
          - type: map_at_5
            value: 11.125
          - type: mrr_at_1
            value: 48.75
          - type: mrr_at_10
            value: 58.425000000000004
          - type: mrr_at_100
            value: 59.075
          - type: mrr_at_1000
            value: 59.095
          - type: mrr_at_3
            value: 56.291999999999994
          - type: mrr_at_5
            value: 57.679
          - type: ndcg_at_1
            value: 37.875
          - type: ndcg_at_10
            value: 27.77
          - type: ndcg_at_100
            value: 30.288999999999998
          - type: ndcg_at_1000
            value: 36.187999999999995
          - type: ndcg_at_3
            value: 31.385999999999996
          - type: ndcg_at_5
            value: 29.923
          - type: precision_at_1
            value: 48.75
          - type: precision_at_10
            value: 22.375
          - type: precision_at_100
            value: 6.3420000000000005
          - type: precision_at_1000
            value: 1.4489999999999998
          - type: precision_at_3
            value: 35.5
          - type: precision_at_5
            value: 30.55
          - type: recall_at_1
            value: 6.343
          - type: recall_at_10
            value: 16.936
          - type: recall_at_100
            value: 35.955999999999996
          - type: recall_at_1000
            value: 55.787
          - type: recall_at_3
            value: 10.771
          - type: recall_at_5
            value: 13.669999999999998
      - task:
          type: Classification
        dataset:
          type: mteb/emotion
          name: MTEB EmotionClassification
          config: default
          split: test
          revision: 4f58c6b202a23cf9a4da393831edf4f9183cad37
        metrics:
          - type: accuracy
            value: 41.99
          - type: f1
            value: 36.823402174564954
      - task:
          type: Retrieval
        dataset:
          type: fever
          name: MTEB FEVER
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 40.088
          - type: map_at_10
            value: 52.69200000000001
          - type: map_at_100
            value: 53.296
          - type: map_at_1000
            value: 53.325
          - type: map_at_3
            value: 49.905
          - type: map_at_5
            value: 51.617000000000004
          - type: mrr_at_1
            value: 43.009
          - type: mrr_at_10
            value: 56.203
          - type: mrr_at_100
            value: 56.75
          - type: mrr_at_1000
            value: 56.769000000000005
          - type: mrr_at_3
            value: 53.400000000000006
          - type: mrr_at_5
            value: 55.163
          - type: ndcg_at_1
            value: 43.009
          - type: ndcg_at_10
            value: 59.39
          - type: ndcg_at_100
            value: 62.129999999999995
          - type: ndcg_at_1000
            value: 62.793
          - type: ndcg_at_3
            value: 53.878
          - type: ndcg_at_5
            value: 56.887
          - type: precision_at_1
            value: 43.009
          - type: precision_at_10
            value: 8.366
          - type: precision_at_100
            value: 0.983
          - type: precision_at_1000
            value: 0.105
          - type: precision_at_3
            value: 22.377
          - type: precision_at_5
            value: 15.035000000000002
          - type: recall_at_1
            value: 40.088
          - type: recall_at_10
            value: 76.68700000000001
          - type: recall_at_100
            value: 88.91
          - type: recall_at_1000
            value: 93.782
          - type: recall_at_3
            value: 61.809999999999995
          - type: recall_at_5
            value: 69.131
      - task:
          type: Retrieval
        dataset:
          type: fiqa
          name: MTEB FiQA2018
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 10.817
          - type: map_at_10
            value: 18.9
          - type: map_at_100
            value: 20.448
          - type: map_at_1000
            value: 20.660999999999998
          - type: map_at_3
            value: 15.979
          - type: map_at_5
            value: 17.415
          - type: mrr_at_1
            value: 23.148
          - type: mrr_at_10
            value: 31.208000000000002
          - type: mrr_at_100
            value: 32.167
          - type: mrr_at_1000
            value: 32.242
          - type: mrr_at_3
            value: 28.498
          - type: mrr_at_5
            value: 29.964000000000002
          - type: ndcg_at_1
            value: 23.148
          - type: ndcg_at_10
            value: 25.325999999999997
          - type: ndcg_at_100
            value: 31.927
          - type: ndcg_at_1000
            value: 36.081
          - type: ndcg_at_3
            value: 21.647
          - type: ndcg_at_5
            value: 22.762999999999998
          - type: precision_at_1
            value: 23.148
          - type: precision_at_10
            value: 7.546
          - type: precision_at_100
            value: 1.415
          - type: precision_at_1000
            value: 0.216
          - type: precision_at_3
            value: 14.969
          - type: precision_at_5
            value: 11.327
          - type: recall_at_1
            value: 10.817
          - type: recall_at_10
            value: 32.164
          - type: recall_at_100
            value: 57.655
          - type: recall_at_1000
            value: 82.797
          - type: recall_at_3
            value: 19.709
          - type: recall_at_5
            value: 24.333
      - task:
          type: Retrieval
        dataset:
          type: hotpotqa
          name: MTEB HotpotQA
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 25.380999999999997
          - type: map_at_10
            value: 33.14
          - type: map_at_100
            value: 33.948
          - type: map_at_1000
            value: 34.028000000000006
          - type: map_at_3
            value: 31.019999999999996
          - type: map_at_5
            value: 32.23
          - type: mrr_at_1
            value: 50.763000000000005
          - type: mrr_at_10
            value: 57.899
          - type: mrr_at_100
            value: 58.426
          - type: mrr_at_1000
            value: 58.457
          - type: mrr_at_3
            value: 56.093
          - type: mrr_at_5
            value: 57.116
          - type: ndcg_at_1
            value: 50.763000000000005
          - type: ndcg_at_10
            value: 41.656
          - type: ndcg_at_100
            value: 45.079
          - type: ndcg_at_1000
            value: 46.916999999999994
          - type: ndcg_at_3
            value: 37.834
          - type: ndcg_at_5
            value: 39.732
          - type: precision_at_1
            value: 50.763000000000005
          - type: precision_at_10
            value: 8.648
          - type: precision_at_100
            value: 1.135
          - type: precision_at_1000
            value: 0.13799999999999998
          - type: precision_at_3
            value: 23.105999999999998
          - type: precision_at_5
            value: 15.363
          - type: recall_at_1
            value: 25.380999999999997
          - type: recall_at_10
            value: 43.241
          - type: recall_at_100
            value: 56.745000000000005
          - type: recall_at_1000
            value: 69.048
          - type: recall_at_3
            value: 34.659
          - type: recall_at_5
            value: 38.406
      - task:
          type: Classification
        dataset:
          type: mteb/imdb
          name: MTEB ImdbClassification
          config: default
          split: test
          revision: 3d86128a09e091d6018b6d26cad27f2739fc2db7
        metrics:
          - type: accuracy
            value: 79.544
          - type: ap
            value: 73.82920133396664
          - type: f1
            value: 79.51048124883265
      - task:
          type: Retrieval
        dataset:
          type: msmarco
          name: MTEB MSMARCO
          config: default
          split: dev
          revision: None
        metrics:
          - type: map_at_1
            value: 11.174000000000001
          - type: map_at_10
            value: 19.451999999999998
          - type: map_at_100
            value: 20.612
          - type: map_at_1000
            value: 20.703
          - type: map_at_3
            value: 16.444
          - type: map_at_5
            value: 18.083
          - type: mrr_at_1
            value: 11.447000000000001
          - type: mrr_at_10
            value: 19.808
          - type: mrr_at_100
            value: 20.958
          - type: mrr_at_1000
            value: 21.041999999999998
          - type: mrr_at_3
            value: 16.791
          - type: mrr_at_5
            value: 18.459
          - type: ndcg_at_1
            value: 11.447000000000001
          - type: ndcg_at_10
            value: 24.556
          - type: ndcg_at_100
            value: 30.637999999999998
          - type: ndcg_at_1000
            value: 33.14
          - type: ndcg_at_3
            value: 18.325
          - type: ndcg_at_5
            value: 21.278
          - type: precision_at_1
            value: 11.447000000000001
          - type: precision_at_10
            value: 4.215
          - type: precision_at_100
            value: 0.732
          - type: precision_at_1000
            value: 0.095
          - type: precision_at_3
            value: 8.052
          - type: precision_at_5
            value: 6.318
          - type: recall_at_1
            value: 11.174000000000001
          - type: recall_at_10
            value: 40.543
          - type: recall_at_100
            value: 69.699
          - type: recall_at_1000
            value: 89.403
          - type: recall_at_3
            value: 23.442
          - type: recall_at_5
            value: 30.536
      - task:
          type: Classification
        dataset:
          type: mteb/mtop_domain
          name: MTEB MTOPDomainClassification (en)
          config: en
          split: test
          revision: d80d48c1eb48d3562165c59d59d0034df9fff0bf
        metrics:
          - type: accuracy
            value: 89.6671226630187
          - type: f1
            value: 89.57660424361246
      - task:
          type: Classification
        dataset:
          type: mteb/mtop_intent
          name: MTEB MTOPIntentClassification (en)
          config: en
          split: test
          revision: ae001d0e6b1228650b7bd1c2c65fb50ad11a8aba
        metrics:
          - type: accuracy
            value: 60.284997720018254
          - type: f1
            value: 40.30637400152823
      - task:
          type: Classification
        dataset:
          type: mteb/amazon_massive_intent
          name: MTEB MassiveIntentClassification (en)
          config: en
          split: test
          revision: 31efe3c427b0bae9c22cbb560b8f15491cc6bed7
        metrics:
          - type: accuracy
            value: 63.33557498318763
          - type: f1
            value: 60.24039910680179
      - task:
          type: Classification
        dataset:
          type: mteb/amazon_massive_scenario
          name: MTEB MassiveScenarioClassification (en)
          config: en
          split: test
          revision: 7d571f92784cd94a019292a1f45445077d0ef634
        metrics:
          - type: accuracy
            value: 72.37390719569603
          - type: f1
            value: 72.33097333477316
      - task:
          type: Clustering
        dataset:
          type: mteb/medrxiv-clustering-p2p
          name: MTEB MedrxivClusteringP2P
          config: default
          split: test
          revision: e7a26af6f3ae46b30dde8737f02c07b1505bcc73
        metrics:
          - type: v_measure
            value: 34.68158939060552
      - task:
          type: Clustering
        dataset:
          type: mteb/medrxiv-clustering-s2s
          name: MTEB MedrxivClusteringS2S
          config: default
          split: test
          revision: 35191c8c0dca72d8ff3efcd72aa802307d469663
        metrics:
          - type: v_measure
            value: 30.340061711905236
      - task:
          type: Reranking
        dataset:
          type: mteb/mind_small
          name: MTEB MindSmallReranking
          config: default
          split: test
          revision: 3bdac13927fdc888b903db93b2ffdbd90b295a69
        metrics:
          - type: map
            value: 32.01814326295803
          - type: mrr
            value: 33.20555240055367
      - task:
          type: Retrieval
        dataset:
          type: nfcorpus
          name: MTEB NFCorpus
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 3.3910000000000005
          - type: map_at_10
            value: 7.7219999999999995
          - type: map_at_100
            value: 10.286
          - type: map_at_1000
            value: 11.668000000000001
          - type: map_at_3
            value: 5.552
          - type: map_at_5
            value: 6.468
          - type: mrr_at_1
            value: 34.365
          - type: mrr_at_10
            value: 42.555
          - type: mrr_at_100
            value: 43.295
          - type: mrr_at_1000
            value: 43.357
          - type: mrr_at_3
            value: 40.299
          - type: mrr_at_5
            value: 41.182
          - type: ndcg_at_1
            value: 31.424000000000003
          - type: ndcg_at_10
            value: 24.758
          - type: ndcg_at_100
            value: 23.677999999999997
          - type: ndcg_at_1000
            value: 33.377
          - type: ndcg_at_3
            value: 28.302
          - type: ndcg_at_5
            value: 26.342
          - type: precision_at_1
            value: 33.437
          - type: precision_at_10
            value: 19.256999999999998
          - type: precision_at_100
            value: 6.662999999999999
          - type: precision_at_1000
            value: 1.9900000000000002
          - type: precision_at_3
            value: 27.761000000000003
          - type: precision_at_5
            value: 23.715
          - type: recall_at_1
            value: 3.3910000000000005
          - type: recall_at_10
            value: 11.068
          - type: recall_at_100
            value: 25.878
          - type: recall_at_1000
            value: 60.19
          - type: recall_at_3
            value: 6.1690000000000005
          - type: recall_at_5
            value: 7.767
      - task:
          type: Retrieval
        dataset:
          type: nq
          name: MTEB NQ
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 15.168000000000001
          - type: map_at_10
            value: 26.177
          - type: map_at_100
            value: 27.564
          - type: map_at_1000
            value: 27.628999999999998
          - type: map_at_3
            value: 22.03
          - type: map_at_5
            value: 24.276
          - type: mrr_at_1
            value: 17.439
          - type: mrr_at_10
            value: 28.205000000000002
          - type: mrr_at_100
            value: 29.357
          - type: mrr_at_1000
            value: 29.408
          - type: mrr_at_3
            value: 24.377
          - type: mrr_at_5
            value: 26.540000000000003
          - type: ndcg_at_1
            value: 17.41
          - type: ndcg_at_10
            value: 32.936
          - type: ndcg_at_100
            value: 39.196999999999996
          - type: ndcg_at_1000
            value: 40.892
          - type: ndcg_at_3
            value: 24.721
          - type: ndcg_at_5
            value: 28.615000000000002
          - type: precision_at_1
            value: 17.41
          - type: precision_at_10
            value: 6.199000000000001
          - type: precision_at_100
            value: 0.9690000000000001
          - type: precision_at_1000
            value: 0.11299999999999999
          - type: precision_at_3
            value: 11.790000000000001
          - type: precision_at_5
            value: 9.264
          - type: recall_at_1
            value: 15.168000000000001
          - type: recall_at_10
            value: 51.914
          - type: recall_at_100
            value: 79.804
          - type: recall_at_1000
            value: 92.75999999999999
          - type: recall_at_3
            value: 30.212
          - type: recall_at_5
            value: 39.204
      - task:
          type: Retrieval
        dataset:
          type: quora
          name: MTEB QuoraRetrieval
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 67.306
          - type: map_at_10
            value: 80.634
          - type: map_at_100
            value: 81.349
          - type: map_at_1000
            value: 81.37299999999999
          - type: map_at_3
            value: 77.691
          - type: map_at_5
            value: 79.512
          - type: mrr_at_1
            value: 77.56
          - type: mrr_at_10
            value: 84.177
          - type: mrr_at_100
            value: 84.35000000000001
          - type: mrr_at_1000
            value: 84.353
          - type: mrr_at_3
            value: 83.003
          - type: mrr_at_5
            value: 83.799
          - type: ndcg_at_1
            value: 77.58
          - type: ndcg_at_10
            value: 84.782
          - type: ndcg_at_100
            value: 86.443
          - type: ndcg_at_1000
            value: 86.654
          - type: ndcg_at_3
            value: 81.67
          - type: ndcg_at_5
            value: 83.356
          - type: precision_at_1
            value: 77.58
          - type: precision_at_10
            value: 12.875
          - type: precision_at_100
            value: 1.503
          - type: precision_at_1000
            value: 0.156
          - type: precision_at_3
            value: 35.63
          - type: precision_at_5
            value: 23.483999999999998
          - type: recall_at_1
            value: 67.306
          - type: recall_at_10
            value: 92.64
          - type: recall_at_100
            value: 98.681
          - type: recall_at_1000
            value: 99.79
          - type: recall_at_3
            value: 83.682
          - type: recall_at_5
            value: 88.424
      - task:
          type: Clustering
        dataset:
          type: mteb/reddit-clustering
          name: MTEB RedditClustering
          config: default
          split: test
          revision: 24640382cdbf8abc73003fb0fa6d111a705499eb
        metrics:
          - type: v_measure
            value: 50.76319866126382
      - task:
          type: Clustering
        dataset:
          type: mteb/reddit-clustering-p2p
          name: MTEB RedditClusteringP2P
          config: default
          split: test
          revision: 282350215ef01743dc01b456c7f5241fa8937f16
        metrics:
          - type: v_measure
            value: 55.024711941648995
      - task:
          type: Retrieval
        dataset:
          type: scidocs
          name: MTEB SCIDOCS
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 3.9379999999999997
          - type: map_at_10
            value: 8.817
          - type: map_at_100
            value: 10.546999999999999
          - type: map_at_1000
            value: 10.852
          - type: map_at_3
            value: 6.351999999999999
          - type: map_at_5
            value: 7.453
          - type: mrr_at_1
            value: 19.400000000000002
          - type: mrr_at_10
            value: 27.371000000000002
          - type: mrr_at_100
            value: 28.671999999999997
          - type: mrr_at_1000
            value: 28.747
          - type: mrr_at_3
            value: 24.583
          - type: mrr_at_5
            value: 26.143
          - type: ndcg_at_1
            value: 19.400000000000002
          - type: ndcg_at_10
            value: 15.264
          - type: ndcg_at_100
            value: 22.63
          - type: ndcg_at_1000
            value: 28.559
          - type: ndcg_at_3
            value: 14.424999999999999
          - type: ndcg_at_5
            value: 12.520000000000001
          - type: precision_at_1
            value: 19.400000000000002
          - type: precision_at_10
            value: 7.8100000000000005
          - type: precision_at_100
            value: 1.854
          - type: precision_at_1000
            value: 0.329
          - type: precision_at_3
            value: 13.100000000000001
          - type: precision_at_5
            value: 10.68
          - type: recall_at_1
            value: 3.9379999999999997
          - type: recall_at_10
            value: 15.903
          - type: recall_at_100
            value: 37.645
          - type: recall_at_1000
            value: 66.86
          - type: recall_at_3
            value: 7.993
          - type: recall_at_5
            value: 10.885
      - task:
          type: STS
        dataset:
          type: mteb/sickr-sts
          name: MTEB SICK-R
          config: default
          split: test
          revision: a6ea5a8cab320b040a23452cc28066d9beae2cee
        metrics:
          - type: cos_sim_pearson
            value: 80.12689060151425
          - type: cos_sim_spearman
            value: 70.46515535094771
          - type: euclidean_pearson
            value: 77.17160003557223
          - type: euclidean_spearman
            value: 70.4651757047438
          - type: manhattan_pearson
            value: 77.18129609281937
          - type: manhattan_spearman
            value: 70.46610403752913
      - task:
          type: STS
        dataset:
          type: mteb/sts12-sts
          name: MTEB STS12
          config: default
          split: test
          revision: a0d554a64d88156834ff5ae9920b964011b16384
        metrics:
          - type: cos_sim_pearson
            value: 70.451157033355
          - type: cos_sim_spearman
            value: 63.99899601697852
          - type: euclidean_pearson
            value: 67.46985359967678
          - type: euclidean_spearman
            value: 64.00001637764805
          - type: manhattan_pearson
            value: 67.56534741780037
          - type: manhattan_spearman
            value: 64.06533893575366
      - task:
          type: STS
        dataset:
          type: mteb/sts13-sts
          name: MTEB STS13
          config: default
          split: test
          revision: 7e90230a92c190f1bf69ae9002b8cea547a64cca
        metrics:
          - type: cos_sim_pearson
            value: 77.65086614464292
          - type: cos_sim_spearman
            value: 78.20169706921848
          - type: euclidean_pearson
            value: 77.77758172155283
          - type: euclidean_spearman
            value: 78.20169706921848
          - type: manhattan_pearson
            value: 77.75077884860052
          - type: manhattan_spearman
            value: 78.16875216484164
      - task:
          type: STS
        dataset:
          type: mteb/sts14-sts
          name: MTEB STS14
          config: default
          split: test
          revision: 6031580fec1f6af667f0bd2da0a551cf4f0b2375
        metrics:
          - type: cos_sim_pearson
            value: 76.26381598259717
          - type: cos_sim_spearman
            value: 70.78377709313477
          - type: euclidean_pearson
            value: 74.82646556532096
          - type: euclidean_spearman
            value: 70.78377658155212
          - type: manhattan_pearson
            value: 74.81784766108225
          - type: manhattan_spearman
            value: 70.79351454692176
      - task:
          type: STS
        dataset:
          type: mteb/sts15-sts
          name: MTEB STS15
          config: default
          split: test
          revision: ae752c7c21bf194d8b67fd573edf7ae58183cbe3
        metrics:
          - type: cos_sim_pearson
            value: 79.00532026789739
          - type: cos_sim_spearman
            value: 80.02708383244838
          - type: euclidean_pearson
            value: 79.48345422610525
          - type: euclidean_spearman
            value: 80.02708383244838
          - type: manhattan_pearson
            value: 79.44519739854803
          - type: manhattan_spearman
            value: 79.98344094559687
      - task:
          type: STS
        dataset:
          type: mteb/sts16-sts
          name: MTEB STS16
          config: default
          split: test
          revision: 4d8694f8f0e0100860b497b999b3dbed754a0513
        metrics:
          - type: cos_sim_pearson
            value: 77.32783048164805
          - type: cos_sim_spearman
            value: 78.79729961288045
          - type: euclidean_pearson
            value: 78.72111945793154
          - type: euclidean_spearman
            value: 78.79729904606872
          - type: manhattan_pearson
            value: 78.72464311117116
          - type: manhattan_spearman
            value: 78.822591248334
      - task:
          type: STS
        dataset:
          type: mteb/sts17-crosslingual-sts
          name: MTEB STS17 (en-en)
          config: en-en
          split: test
          revision: af5e6fb845001ecf41f4c1e033ce921939a2a68d
        metrics:
          - type: cos_sim_pearson
            value: 82.04318630630854
          - type: cos_sim_spearman
            value: 83.87886389259836
          - type: euclidean_pearson
            value: 83.40385877895086
          - type: euclidean_spearman
            value: 83.87886389259836
          - type: manhattan_pearson
            value: 83.46337128901547
          - type: manhattan_spearman
            value: 83.9723106941644
      - task:
          type: STS
        dataset:
          type: mteb/sts22-crosslingual-sts
          name: MTEB STS22 (en)
          config: en
          split: test
          revision: 6d1ba47164174a496b7fa5d3569dae26a6813b80
        metrics:
          - type: cos_sim_pearson
            value: 63.003511169944595
          - type: cos_sim_spearman
            value: 64.39318805580227
          - type: euclidean_pearson
            value: 65.4797990735967
          - type: euclidean_spearman
            value: 64.39318805580227
          - type: manhattan_pearson
            value: 65.44604544280844
          - type: manhattan_spearman
            value: 64.38742899984233
      - task:
          type: STS
        dataset:
          type: mteb/stsbenchmark-sts
          name: MTEB STSBenchmark
          config: default
          split: test
          revision: b0fddb56ed78048fa8b90373c8a3cfc37b684831
        metrics:
          - type: cos_sim_pearson
            value: 76.63101237585029
          - type: cos_sim_spearman
            value: 75.57446967644269
          - type: euclidean_pearson
            value: 76.93491768734478
          - type: euclidean_spearman
            value: 75.57446967644269
          - type: manhattan_pearson
            value: 76.92187567800636
          - type: manhattan_spearman
            value: 75.57239337194585
      - task:
          type: Reranking
        dataset:
          type: mteb/scidocs-reranking
          name: MTEB SciDocsRR
          config: default
          split: test
          revision: d3c5e1fc0b855ab6097bf1cda04dd73947d7caab
        metrics:
          - type: map
            value: 78.5376604868993
          - type: mrr
            value: 92.94422897364073
      - task:
          type: Retrieval
        dataset:
          type: scifact
          name: MTEB SciFact
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 38.872
          - type: map_at_10
            value: 50.417
          - type: map_at_100
            value: 51.202000000000005
          - type: map_at_1000
            value: 51.25999999999999
          - type: map_at_3
            value: 47.02
          - type: map_at_5
            value: 49.326
          - type: mrr_at_1
            value: 41
          - type: mrr_at_10
            value: 51.674
          - type: mrr_at_100
            value: 52.32599999999999
          - type: mrr_at_1000
            value: 52.376999999999995
          - type: mrr_at_3
            value: 48.778
          - type: mrr_at_5
            value: 50.744
          - type: ndcg_at_1
            value: 41
          - type: ndcg_at_10
            value: 56.027
          - type: ndcg_at_100
            value: 59.362
          - type: ndcg_at_1000
            value: 60.839
          - type: ndcg_at_3
            value: 50.019999999999996
          - type: ndcg_at_5
            value: 53.644999999999996
          - type: precision_at_1
            value: 41
          - type: precision_at_10
            value: 8.1
          - type: precision_at_100
            value: 0.987
          - type: precision_at_1000
            value: 0.11100000000000002
          - type: precision_at_3
            value: 20.444000000000003
          - type: precision_at_5
            value: 14.466999999999999
          - type: recall_at_1
            value: 38.872
          - type: recall_at_10
            value: 71.906
          - type: recall_at_100
            value: 86.367
          - type: recall_at_1000
            value: 98
          - type: recall_at_3
            value: 56.206
          - type: recall_at_5
            value: 65.05
      - task:
          type: PairClassification
        dataset:
          type: mteb/sprintduplicatequestions-pairclassification
          name: MTEB SprintDuplicateQuestions
          config: default
          split: test
          revision: d66bd1f72af766a5cc4b0ca5e00c162f89e8cc46
        metrics:
          - type: cos_sim_accuracy
            value: 99.7039603960396
          - type: cos_sim_ap
            value: 90.40809844250262
          - type: cos_sim_f1
            value: 84.53181583031557
          - type: cos_sim_precision
            value: 87.56698821007502
          - type: cos_sim_recall
            value: 81.69999999999999
          - type: dot_accuracy
            value: 99.7039603960396
          - type: dot_ap
            value: 90.40809844250262
          - type: dot_f1
            value: 84.53181583031557
          - type: dot_precision
            value: 87.56698821007502
          - type: dot_recall
            value: 81.69999999999999
          - type: euclidean_accuracy
            value: 99.7039603960396
          - type: euclidean_ap
            value: 90.4080982863383
          - type: euclidean_f1
            value: 84.53181583031557
          - type: euclidean_precision
            value: 87.56698821007502
          - type: euclidean_recall
            value: 81.69999999999999
          - type: manhattan_accuracy
            value: 99.7
          - type: manhattan_ap
            value: 90.39771161966652
          - type: manhattan_f1
            value: 84.32989690721648
          - type: manhattan_precision
            value: 87.02127659574468
          - type: manhattan_recall
            value: 81.8
          - type: max_accuracy
            value: 99.7039603960396
          - type: max_ap
            value: 90.40809844250262
          - type: max_f1
            value: 84.53181583031557
      - task:
          type: Clustering
        dataset:
          type: mteb/stackexchange-clustering
          name: MTEB StackExchangeClustering
          config: default
          split: test
          revision: 6cbc1f7b2bc0622f2e39d2c77fa502909748c259
        metrics:
          - type: v_measure
            value: 59.663210666678715
      - task:
          type: Clustering
        dataset:
          type: mteb/stackexchange-clustering-p2p
          name: MTEB StackExchangeClusteringP2P
          config: default
          split: test
          revision: 815ca46b2622cec33ccafc3735d572c266efdb44
        metrics:
          - type: v_measure
            value: 32.107791216468776
      - task:
          type: Reranking
        dataset:
          type: mteb/stackoverflowdupquestions-reranking
          name: MTEB StackOverflowDupQuestions
          config: default
          split: test
          revision: e185fbe320c72810689fc5848eb6114e1ef5ec69
        metrics:
          - type: map
            value: 46.440691925067604
          - type: mrr
            value: 47.03390257618199
      - task:
          type: Summarization
        dataset:
          type: mteb/summeval
          name: MTEB SummEval
          config: default
          split: test
          revision: cda12ad7615edc362dbf25a00fdd61d3b1eaf93c
        metrics:
          - type: cos_sim_pearson
            value: 31.067177519784074
          - type: cos_sim_spearman
            value: 31.234728424648967
          - type: dot_pearson
            value: 31.06717083018107
          - type: dot_spearman
            value: 31.234728424648967
      - task:
          type: Retrieval
        dataset:
          type: trec-covid
          name: MTEB TRECCOVID
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 0.136
          - type: map_at_10
            value: 0.767
          - type: map_at_100
            value: 3.3689999999999998
          - type: map_at_1000
            value: 8.613999999999999
          - type: map_at_3
            value: 0.369
          - type: map_at_5
            value: 0.514
          - type: mrr_at_1
            value: 48
          - type: mrr_at_10
            value: 63.908
          - type: mrr_at_100
            value: 64.615
          - type: mrr_at_1000
            value: 64.615
          - type: mrr_at_3
            value: 62
          - type: mrr_at_5
            value: 63.4
          - type: ndcg_at_1
            value: 44
          - type: ndcg_at_10
            value: 38.579
          - type: ndcg_at_100
            value: 26.409
          - type: ndcg_at_1000
            value: 26.858999999999998
          - type: ndcg_at_3
            value: 47.134
          - type: ndcg_at_5
            value: 43.287
          - type: precision_at_1
            value: 48
          - type: precision_at_10
            value: 40.400000000000006
          - type: precision_at_100
            value: 26.640000000000004
          - type: precision_at_1000
            value: 12.04
          - type: precision_at_3
            value: 52.666999999999994
          - type: precision_at_5
            value: 46.800000000000004
          - type: recall_at_1
            value: 0.136
          - type: recall_at_10
            value: 1.0070000000000001
          - type: recall_at_100
            value: 6.318
          - type: recall_at_1000
            value: 26.522000000000002
          - type: recall_at_3
            value: 0.41700000000000004
          - type: recall_at_5
            value: 0.606
      - task:
          type: Retrieval
        dataset:
          type: webis-touche2020
          name: MTEB Touche2020
          config: default
          split: test
          revision: None
        metrics:
          - type: map_at_1
            value: 1.9949999999999999
          - type: map_at_10
            value: 8.304
          - type: map_at_100
            value: 13.644
          - type: map_at_1000
            value: 15.43
          - type: map_at_3
            value: 4.788
          - type: map_at_5
            value: 6.22
          - type: mrr_at_1
            value: 22.448999999999998
          - type: mrr_at_10
            value: 37.658
          - type: mrr_at_100
            value: 38.491
          - type: mrr_at_1000
            value: 38.503
          - type: mrr_at_3
            value: 32.312999999999995
          - type: mrr_at_5
            value: 35.68
          - type: ndcg_at_1
            value: 21.429000000000002
          - type: ndcg_at_10
            value: 18.995
          - type: ndcg_at_100
            value: 32.029999999999994
          - type: ndcg_at_1000
            value: 44.852
          - type: ndcg_at_3
            value: 19.464000000000002
          - type: ndcg_at_5
            value: 19.172
          - type: precision_at_1
            value: 22.448999999999998
          - type: precision_at_10
            value: 17.143
          - type: precision_at_100
            value: 6.877999999999999
          - type: precision_at_1000
            value: 1.524
          - type: precision_at_3
            value: 21.769
          - type: precision_at_5
            value: 20
          - type: recall_at_1
            value: 1.9949999999999999
          - type: recall_at_10
            value: 13.395999999999999
          - type: recall_at_100
            value: 44.348
          - type: recall_at_1000
            value: 82.622
          - type: recall_at_3
            value: 5.896
          - type: recall_at_5
            value: 8.554
      - task:
          type: Classification
        dataset:
          type: mteb/toxic_conversations_50k
          name: MTEB ToxicConversationsClassification
          config: default
          split: test
          revision: d7c0de2777da35d6aae2200a62c6e0e5af397c4c
        metrics:
          - type: accuracy
            value: 67.9394
          - type: ap
            value: 12.943337263423334
          - type: f1
            value: 52.28243093094156
      - task:
          type: Classification
        dataset:
          type: mteb/tweet_sentiment_extraction
          name: MTEB TweetSentimentExtractionClassification
          config: default
          split: test
          revision: d604517c81ca91fe16a244d1248fc021f9ecee7a
        metrics:
          - type: accuracy
            value: 56.414827391058296
          - type: f1
            value: 56.666412409573105
      - task:
          type: Clustering
        dataset:
          type: mteb/twentynewsgroups-clustering
          name: MTEB TwentyNewsgroupsClustering
          config: default
          split: test
          revision: 6125ec4e24fa026cec8a478383ee943acfbd5449
        metrics:
          - type: v_measure
            value: 47.009746255495465
      - task:
          type: PairClassification
        dataset:
          type: mteb/twittersemeval2015-pairclassification
          name: MTEB TwitterSemEval2015
          config: default
          split: test
          revision: 70970daeab8776df92f5ea462b6173c0b46fd2d1
        metrics:
          - type: cos_sim_accuracy
            value: 84.02574953805807
          - type: cos_sim_ap
            value: 67.66599910763128
          - type: cos_sim_f1
            value: 63.491277990844985
          - type: cos_sim_precision
            value: 59.77172140694154
          - type: cos_sim_recall
            value: 67.70448548812665
          - type: dot_accuracy
            value: 84.02574953805807
          - type: dot_ap
            value: 67.66600090945406
          - type: dot_f1
            value: 63.491277990844985
          - type: dot_precision
            value: 59.77172140694154
          - type: dot_recall
            value: 67.70448548812665
          - type: euclidean_accuracy
            value: 84.02574953805807
          - type: euclidean_ap
            value: 67.6659842364448
          - type: euclidean_f1
            value: 63.491277990844985
          - type: euclidean_precision
            value: 59.77172140694154
          - type: euclidean_recall
            value: 67.70448548812665
          - type: manhattan_accuracy
            value: 84.0317100792752
          - type: manhattan_ap
            value: 67.66351692448987
          - type: manhattan_f1
            value: 63.48610948306178
          - type: manhattan_precision
            value: 57.11875131828729
          - type: manhattan_recall
            value: 71.45118733509234
          - type: max_accuracy
            value: 84.0317100792752
          - type: max_ap
            value: 67.66600090945406
          - type: max_f1
            value: 63.491277990844985
      - task:
          type: PairClassification
        dataset:
          type: mteb/twitterurlcorpus-pairclassification
          name: MTEB TwitterURLCorpus
          config: default
          split: test
          revision: 8b6510b0b1fa4e4c4f879467980e9be563ec1cdf
        metrics:
          - type: cos_sim_accuracy
            value: 87.53832421314084
          - type: cos_sim_ap
            value: 83.11416594316626
          - type: cos_sim_f1
            value: 75.41118114347518
          - type: cos_sim_precision
            value: 73.12839059674504
          - type: cos_sim_recall
            value: 77.8410840776101
          - type: dot_accuracy
            value: 87.53832421314084
          - type: dot_ap
            value: 83.11416226342155
          - type: dot_f1
            value: 75.41118114347518
          - type: dot_precision
            value: 73.12839059674504
          - type: dot_recall
            value: 77.8410840776101
          - type: euclidean_accuracy
            value: 87.53832421314084
          - type: euclidean_ap
            value: 83.11416284455395
          - type: euclidean_f1
            value: 75.41118114347518
          - type: euclidean_precision
            value: 73.12839059674504
          - type: euclidean_recall
            value: 77.8410840776101
          - type: manhattan_accuracy
            value: 87.49369348391353
          - type: manhattan_ap
            value: 83.08066812574694
          - type: manhattan_f1
            value: 75.36561228603892
          - type: manhattan_precision
            value: 71.9202518363064
          - type: manhattan_recall
            value: 79.15768401601478
          - type: max_accuracy
            value: 87.53832421314084
          - type: max_ap
            value: 83.11416594316626
          - type: max_f1
            value: 75.41118114347518

lodestone-base-4096-v1

Hum-Works/lodestone-base-4096-v1. Griffin McCauley, Will Fortin, Dylan DiGioia 2023

This new sentence-transformers model from Hum maps long sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.

Abstract

In the hopes of furthering Hum's overarching mission of increasing the accessibility and interconnectivity of human knowledge, this model was developed as part of a project intending to boost the maximum input sequence length of sentence embedding models by leveraging recent architectural advances in the design of transformer models such as the incorporation of FlashAttention, Attention with Linear Biases (ALiBi), and Gated Linear Units (GLU). These modifications and enhancements were implemented by the team at MosaicML who designed and constructed the pre-trained mosaic-bert-base-seqlen-2048 model, and more information regarding the details of their development and testing specifications can be found on the model card.

While the fine-tuning procedure followed during the course of this project loosely mirrors that of the of the original Flax-sentence-embeddings team responsible for the creation of many other popular sentence-transformers models (e.g. all-mpnet-base-v2, all-distilroberta-v1, and all-MiniLM-L6-v2), our methodology includes novel techniques for data loading, batch sampling, and model checkpointing intended to improve training efficiency with regards to memory allocation and data storage.

Through combining these well-established and proven fine-tuning practices with novel advances in transformer architectural elements, our lodestone-base-4096-v1 model is able to achieve comparable performance metrics on standard text embedding evaluation benchmarks while also supporting a longer and more robust input sequence length of 4096 while retaining a smaller, more manageable size capable of being run on either a GPU or CPU.

Usage

Using this model becomes relatively easy when you have sentence-transformers installed. At the time of publishing, sentence-transformers does not support remote code which is required for flash-attention used by the model. A fork of the sentence-transformers repository that allows remote code execution is provided for convenience. It can be installed using the following command:

pip install git+https://github.com/Hum-Works/sentence-transformers.git
pip install einops

Then you can use the model like this:

from sentence_transformers import SentenceTransformer

model = SentenceTransformer('Hum-Works/lodestone-base-4096-v1', trust_remote_code=True, revision='v1.0.0')
sentences = ["This is an example sentence", "Each sentence is converted"]
embeddings = model.encode(sentences)
print(embeddings)

Note: The model will use the openAI/Triton implementation of FlashAttention if installed. This is more performant than the fallback, torch implementation. Some platforms and GPUs may not be supported by Triton - up to date compatibility can be found on Triton’s github page.


Background

The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained mosaic-bert-base-seqlen-2048 model and fine-tuned it on a nearly 1.5B sentence pairs dataset. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset.

Intended uses

Our model is intended to be used as a long sentence and paragraph encoder. Given an input text, it outputs a vector containing the semantic information. The sentence vector may be used for information retrieval, clustering, or sentence similarity tasks.

Training procedure

Pre-training

We use the pretrained mosaic-bert-base-seqlen-2048. Please refer to the model card for more detailed information about the pre-training procedure.

Fine-tuning

We fine-tune the model using a contrastive objective. Formally, we compute the dot product of each possible sentence pairing in the batch. We then apply the cross entropy loss by comparing with true pairs.

Hyperparameters

We trained our model on an ml.g5.4xlarge EC2 instance with 1 NVIDIA A10G Tensor Core GPU. We train the model during 1.4 million steps using a batch size of 16. We use a learning rate warm up of 500. The sequence length during training was limited to 2048 tokens. We used the AdamW optimizer with a 2e-5 learning rate and weight decay of 0.01 (i.e. the default parameter values for SentenceTransformer.fit()). The full training script is accessible in this current repository: Training.py.

Model Architecture

By incorporating FlashAttention, Attention with Linear Biases (ALiBi), and Gated Linear Units (GLU), this model is able to handle input sequences of 4096, 8x longer than that supported by most comparable sentence embedding models. The model was trained using a sequence length maximum of 2048, but the final model has a maximum sequence length of 4096. This is accomplished by taking advantage of ALiBi’s positional attention extrapolation which has been shown to allow sequence lengths of 2x the initial trained length.

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 4096, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False})
  (2): Normalize()
)

Training data

We use the concatenation from multiple datasets to fine-tune our model. The total number of sentence pairs is nearly 1.5 billion sentences. We sampled each dataset given a weighted probability proportional to its relative contribution to the entire dataset. The breakdown of the dataset can be seen below, and the entire dataset can be publicly accessed and uploaded via the Dataloading.ipynb located within this repository.

Dataset Paper Number of training tuples
Reddit comments (2015-2018) paper 726,484,430
S2ORC Citation pairs (Abstracts) paper 252,102,397
Reddit posts (Title, Body) pairs - 127,445,911
Amazon reviews (2018) (Title, Review) pairs - 87,877,725
WikiAnswers Duplicate question pairs paper 77,427,422
PAQ (Question, Answer) pairs paper 64,371,441
S2ORC Citation pairs (Titles) paper 52,603,982
S2ORC (Title, Abstract) paper 41,769,185
Stack Exchange (Title, Body) pairs - 25,368,423
MS MARCO triplets paper 9,144,553
Stack Exchange (Title, Most Upvoted Answer) pairs - 4,784,250
Stack Exchange (Title+Body, Most Upvoted Answer) pairs - 4,551,660
GOOAQ: Open Question Answering with Diverse Answer Types paper 3,012,496
Amazon QA - 2,507,114
Code Search - 1,375,067
Yahoo Answers (Title, Answer) paper 1,198,260
[AG News]((Title, Description) pairs of news articles from the AG News dataset) - 1,157,745
COCO Image captions paper 828,395
SPECTER citation triplets paper 684,100
Yahoo Answers (Question, Answer) paper 681,164
Yahoo Answers (Title, Question) paper 659,896
CC News (Title, article) pairs - 614,664
NPR (Title, Body) pairs - 594,384
SearchQA paper 582,261
MS Marco (Query, Answer Passage) pairs paper 532,751
Stack Exchange (Title, Body) pairs - 364,000
Eli5 paper 325,475
Flickr 30k paper 317,695
CNN & DailyMail (highlight sentences, article) pairs - 311,971
Stack Exchange Duplicate questions (titles) - 304,524
AllNLI (SNLI and MultiNLI paper SNLI, paper MultiNLI 277,230
Stack Exchange Duplicate questions (bodies) - 250,518
Stack Exchange Duplicate questions (titles+bodies) - 250,459
XSUM (Summary, News Article) pairs - 226,711
Stack Exchange (Title+Body, Most Upvoted Answer, Most Downvoted Answer) triplets - 216,454
Sentence Compression paper 180,000
FEVER training data - 139,051
Wikihow paper 128,542
SearchQA (Question, Top-Snippet) paper 117,384
Altlex paper 112,696
Quora Question Duplicates - 103,663
Quora Question Triplets - 103,663
Simple Wikipedia paper 102,225
Natural Questions (NQ) paper 100,231
SQuAD2.0 paper 87,599
TriviaQA - 73,346
Total 1,492,453,113

Replication

The entire fine-tuning process for this model can be replicated by following the steps outlined in the Replication.txt file within this repository. This document explains how to modify the sentence-transformers library, configure the pre-trained mosaic-bert-base-seqlen-2048 model, load all of the training data, and execute the training script.

Limitations

Due to technical constraints (e.g. limited GPU memory capacity), this model was trained with a smaller batch size of 16, making it so that each step during training was less well-informed than it would have been on a higher performance system. This smaller than ideal hyperparameter value will generally cause the model to be more likely to get stuck in a local minimum and for the parameter configuration to take a longer time to converge to the optimum. In order to counteract this potential risk, we trained the model for a larger number of steps than many of its contemporaries to ensure a greater chance of achieving strong performance, but this is an area which could be improved if further fine-tuning was performed.

It is also worth noting that, while this model is able to handle longer input sequences of up to 4096 word pieces, the training dataset used consists of sentence and paragraph pairs and triplets which do not necessarily reach that maximum sequence length. Since the data was not tailored specifically for this larger input size, further fine-tuning may be required to ensure highly accurate embeddings for longer texts of that magnitude.

Finally, as stated on https://huggingface.co/datasets/sentence-transformers/reddit-title-body, an additional reminder and warning regarding the Reddit posts data is that one should "Be aware that this dataset is not filtered for biases, hate-speech, spam, racial slurs etc. It depicts the content as it is posted on Reddit." Thus, while we believe this has not induced any pathological behaviors in the model's performance due to its relatively low prevalence of records in the whole dataset of nearly 1.5B sentence pairs and the fact that this model was trained to produce semantic embeddings rather than generative text outputs, it is always important to be aware of vulnerabilities to bias.