---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
- mteb
model-index:
- name: mmlw-e5-base
  results:
  - task:
      type: Clustering
    dataset:
      type: PL-MTEB/8tags-clustering
      name: MTEB 8TagsClustering
      config: default
      split: test
      revision: None
    metrics:
    - type: v_measure
      value: 30.249113010261492
  - task:
      type: Classification
    dataset:
      type: PL-MTEB/allegro-reviews
      name: MTEB AllegroReviews
      config: default
      split: test
      revision: None
    metrics:
    - type: accuracy
      value: 36.3817097415507
    - type: f1
      value: 32.77742158736663
  - task:
      type: Retrieval
    dataset:
      type: arguana-pl
      name: MTEB ArguAna-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 32.646
    - type: map_at_10
      value: 49.488
    - type: map_at_100
      value: 50.190999999999995
    - type: map_at_1000
      value: 50.194
    - type: map_at_3
      value: 44.749
    - type: map_at_5
      value: 47.571999999999996
    - type: mrr_at_1
      value: 34.211000000000006
    - type: mrr_at_10
      value: 50.112
    - type: mrr_at_100
      value: 50.836000000000006
    - type: mrr_at_1000
      value: 50.839
    - type: mrr_at_3
      value: 45.614
    - type: mrr_at_5
      value: 48.242000000000004
    - type: ndcg_at_1
      value: 32.646
    - type: ndcg_at_10
      value: 58.396
    - type: ndcg_at_100
      value: 61.285000000000004
    - type: ndcg_at_1000
      value: 61.358999999999995
    - type: ndcg_at_3
      value: 48.759
    - type: ndcg_at_5
      value: 53.807
    - type: precision_at_1
      value: 32.646
    - type: precision_at_10
      value: 8.663
    - type: precision_at_100
      value: 0.9900000000000001
    - type: precision_at_1000
      value: 0.1
    - type: precision_at_3
      value: 20.128
    - type: precision_at_5
      value: 14.509
    - type: recall_at_1
      value: 32.646
    - type: recall_at_10
      value: 86.629
    - type: recall_at_100
      value: 99.004
    - type: recall_at_1000
      value: 99.57300000000001
    - type: recall_at_3
      value: 60.38400000000001
    - type: recall_at_5
      value: 72.54599999999999
  - task:
      type: Classification
    dataset:
      type: PL-MTEB/cbd
      name: MTEB CBD
      config: default
      split: test
      revision: None
    metrics:
    - type: accuracy
      value: 65.53999999999999
    - type: ap
      value: 19.75395945379771
    - type: f1
      value: 55.00481388401326
  - task:
      type: PairClassification
    dataset:
      type: PL-MTEB/cdsce-pairclassification
      name: MTEB CDSC-E
      config: default
      split: test
      revision: None
    metrics:
    - type: cos_sim_accuracy
      value: 89.5
    - type: cos_sim_ap
      value: 77.26879308078568
    - type: cos_sim_f1
      value: 65.13157894736842
    - type: cos_sim_precision
      value: 86.8421052631579
    - type: cos_sim_recall
      value: 52.10526315789473
    - type: dot_accuracy
      value: 88.0
    - type: dot_ap
      value: 69.17235659054914
    - type: dot_f1
      value: 65.71428571428571
    - type: dot_precision
      value: 71.875
    - type: dot_recall
      value: 60.526315789473685
    - type: euclidean_accuracy
      value: 89.5
    - type: euclidean_ap
      value: 77.1905400565015
    - type: euclidean_f1
      value: 64.91803278688525
    - type: euclidean_precision
      value: 86.08695652173914
    - type: euclidean_recall
      value: 52.10526315789473
    - type: manhattan_accuracy
      value: 89.5
    - type: manhattan_ap
      value: 77.19531778873724
    - type: manhattan_f1
      value: 64.72491909385113
    - type: manhattan_precision
      value: 84.03361344537815
    - type: manhattan_recall
      value: 52.63157894736842
    - type: max_accuracy
      value: 89.5
    - type: max_ap
      value: 77.26879308078568
    - type: max_f1
      value: 65.71428571428571
  - task:
      type: STS
    dataset:
      type: PL-MTEB/cdscr-sts
      name: MTEB CDSC-R
      config: default
      split: test
      revision: None
    metrics:
    - type: cos_sim_pearson
      value: 93.18498922236566
    - type: cos_sim_spearman
      value: 93.26224500108704
    - type: euclidean_pearson
      value: 92.25462061070286
    - type: euclidean_spearman
      value: 93.18623989769242
    - type: manhattan_pearson
      value: 92.16291103586255
    - type: manhattan_spearman
      value: 93.14403078934417
  - task:
      type: Retrieval
    dataset:
      type: dbpedia-pl
      name: MTEB DBPedia-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 8.268
    - type: map_at_10
      value: 17.391000000000002
    - type: map_at_100
      value: 24.266
    - type: map_at_1000
      value: 25.844
    - type: map_at_3
      value: 12.636
    - type: map_at_5
      value: 14.701
    - type: mrr_at_1
      value: 62.74999999999999
    - type: mrr_at_10
      value: 70.25200000000001
    - type: mrr_at_100
      value: 70.601
    - type: mrr_at_1000
      value: 70.613
    - type: mrr_at_3
      value: 68.083
    - type: mrr_at_5
      value: 69.37100000000001
    - type: ndcg_at_1
      value: 51.87500000000001
    - type: ndcg_at_10
      value: 37.185
    - type: ndcg_at_100
      value: 41.949
    - type: ndcg_at_1000
      value: 49.523
    - type: ndcg_at_3
      value: 41.556
    - type: ndcg_at_5
      value: 39.278
    - type: precision_at_1
      value: 63.24999999999999
    - type: precision_at_10
      value: 29.225
    - type: precision_at_100
      value: 9.745
    - type: precision_at_1000
      value: 2.046
    - type: precision_at_3
      value: 43.833
    - type: precision_at_5
      value: 37.9
    - type: recall_at_1
      value: 8.268
    - type: recall_at_10
      value: 22.542
    - type: recall_at_100
      value: 48.154
    - type: recall_at_1000
      value: 72.62100000000001
    - type: recall_at_3
      value: 13.818
    - type: recall_at_5
      value: 17.137
  - task:
      type: Retrieval
    dataset:
      type: fiqa-pl
      name: MTEB FiQA-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 16.489
    - type: map_at_10
      value: 26.916
    - type: map_at_100
      value: 28.582
    - type: map_at_1000
      value: 28.774
    - type: map_at_3
      value: 23.048
    - type: map_at_5
      value: 24.977
    - type: mrr_at_1
      value: 33.642
    - type: mrr_at_10
      value: 41.987
    - type: mrr_at_100
      value: 42.882
    - type: mrr_at_1000
      value: 42.93
    - type: mrr_at_3
      value: 39.48
    - type: mrr_at_5
      value: 40.923
    - type: ndcg_at_1
      value: 33.488
    - type: ndcg_at_10
      value: 34.528
    - type: ndcg_at_100
      value: 41.085
    - type: ndcg_at_1000
      value: 44.474000000000004
    - type: ndcg_at_3
      value: 30.469
    - type: ndcg_at_5
      value: 31.618000000000002
    - type: precision_at_1
      value: 33.488
    - type: precision_at_10
      value: 9.783999999999999
    - type: precision_at_100
      value: 1.6389999999999998
    - type: precision_at_1000
      value: 0.22699999999999998
    - type: precision_at_3
      value: 20.525
    - type: precision_at_5
      value: 15.093
    - type: recall_at_1
      value: 16.489
    - type: recall_at_10
      value: 42.370000000000005
    - type: recall_at_100
      value: 67.183
    - type: recall_at_1000
      value: 87.211
    - type: recall_at_3
      value: 27.689999999999998
    - type: recall_at_5
      value: 33.408
  - task:
      type: Retrieval
    dataset:
      type: hotpotqa-pl
      name: MTEB HotpotQA-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 37.373
    - type: map_at_10
      value: 57.509
    - type: map_at_100
      value: 58.451
    - type: map_at_1000
      value: 58.524
    - type: map_at_3
      value: 54.064
    - type: map_at_5
      value: 56.257999999999996
    - type: mrr_at_1
      value: 74.895
    - type: mrr_at_10
      value: 81.233
    - type: mrr_at_100
      value: 81.461
    - type: mrr_at_1000
      value: 81.47
    - type: mrr_at_3
      value: 80.124
    - type: mrr_at_5
      value: 80.862
    - type: ndcg_at_1
      value: 74.747
    - type: ndcg_at_10
      value: 66.249
    - type: ndcg_at_100
      value: 69.513
    - type: ndcg_at_1000
      value: 70.896
    - type: ndcg_at_3
      value: 61.312
    - type: ndcg_at_5
      value: 64.132
    - type: precision_at_1
      value: 74.747
    - type: precision_at_10
      value: 13.873
    - type: precision_at_100
      value: 1.641
    - type: precision_at_1000
      value: 0.182
    - type: precision_at_3
      value: 38.987
    - type: precision_at_5
      value: 25.621
    - type: recall_at_1
      value: 37.373
    - type: recall_at_10
      value: 69.365
    - type: recall_at_100
      value: 82.039
    - type: recall_at_1000
      value: 91.148
    - type: recall_at_3
      value: 58.48100000000001
    - type: recall_at_5
      value: 64.051
  - task:
      type: Retrieval
    dataset:
      type: msmarco-pl
      name: MTEB MSMARCO-PL
      config: default
      split: validation
      revision: None
    metrics:
    - type: map_at_1
      value: 16.753999999999998
    - type: map_at_10
      value: 26.764
    - type: map_at_100
      value: 27.929
    - type: map_at_1000
      value: 27.994999999999997
    - type: map_at_3
      value: 23.527
    - type: map_at_5
      value: 25.343
    - type: mrr_at_1
      value: 17.192
    - type: mrr_at_10
      value: 27.141
    - type: mrr_at_100
      value: 28.269
    - type: mrr_at_1000
      value: 28.327999999999996
    - type: mrr_at_3
      value: 23.906
    - type: mrr_at_5
      value: 25.759999999999998
    - type: ndcg_at_1
      value: 17.177999999999997
    - type: ndcg_at_10
      value: 32.539
    - type: ndcg_at_100
      value: 38.383
    - type: ndcg_at_1000
      value: 40.132
    - type: ndcg_at_3
      value: 25.884
    - type: ndcg_at_5
      value: 29.15
    - type: precision_at_1
      value: 17.177999999999997
    - type: precision_at_10
      value: 5.268
    - type: precision_at_100
      value: 0.823
    - type: precision_at_1000
      value: 0.097
    - type: precision_at_3
      value: 11.122
    - type: precision_at_5
      value: 8.338
    - type: recall_at_1
      value: 16.753999999999998
    - type: recall_at_10
      value: 50.388
    - type: recall_at_100
      value: 77.86999999999999
    - type: recall_at_1000
      value: 91.55
    - type: recall_at_3
      value: 32.186
    - type: recall_at_5
      value: 40.048
  - task:
      type: Classification
    dataset:
      type: mteb/amazon_massive_intent
      name: MTEB MassiveIntentClassification (pl)
      config: pl
      split: test
      revision: 31efe3c427b0bae9c22cbb560b8f15491cc6bed7
    metrics:
    - type: accuracy
      value: 70.9280430396772
    - type: f1
      value: 68.7099581466286
  - task:
      type: Classification
    dataset:
      type: mteb/amazon_massive_scenario
      name: MTEB MassiveScenarioClassification (pl)
      config: pl
      split: test
      revision: 7d571f92784cd94a019292a1f45445077d0ef634
    metrics:
    - type: accuracy
      value: 74.76126429051783
    - type: f1
      value: 74.72274307018111
  - task:
      type: Retrieval
    dataset:
      type: nfcorpus-pl
      name: MTEB NFCorpus-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 5.348
    - type: map_at_10
      value: 12.277000000000001
    - type: map_at_100
      value: 15.804000000000002
    - type: map_at_1000
      value: 17.277
    - type: map_at_3
      value: 8.783000000000001
    - type: map_at_5
      value: 10.314
    - type: mrr_at_1
      value: 43.963
    - type: mrr_at_10
      value: 52.459999999999994
    - type: mrr_at_100
      value: 53.233
    - type: mrr_at_1000
      value: 53.26499999999999
    - type: mrr_at_3
      value: 50.464
    - type: mrr_at_5
      value: 51.548
    - type: ndcg_at_1
      value: 40.711999999999996
    - type: ndcg_at_10
      value: 33.709
    - type: ndcg_at_100
      value: 31.398
    - type: ndcg_at_1000
      value: 40.042
    - type: ndcg_at_3
      value: 37.85
    - type: ndcg_at_5
      value: 36.260999999999996
    - type: precision_at_1
      value: 43.344
    - type: precision_at_10
      value: 25.851000000000003
    - type: precision_at_100
      value: 8.279
    - type: precision_at_1000
      value: 2.085
    - type: precision_at_3
      value: 36.326
    - type: precision_at_5
      value: 32.074000000000005
    - type: recall_at_1
      value: 5.348
    - type: recall_at_10
      value: 16.441
    - type: recall_at_100
      value: 32.975
    - type: recall_at_1000
      value: 64.357
    - type: recall_at_3
      value: 9.841999999999999
    - type: recall_at_5
      value: 12.463000000000001
  - task:
      type: Retrieval
    dataset:
      type: nq-pl
      name: MTEB NQ-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 24.674
    - type: map_at_10
      value: 37.672
    - type: map_at_100
      value: 38.767
    - type: map_at_1000
      value: 38.82
    - type: map_at_3
      value: 33.823
    - type: map_at_5
      value: 36.063
    - type: mrr_at_1
      value: 27.839000000000002
    - type: mrr_at_10
      value: 40.129
    - type: mrr_at_100
      value: 41.008
    - type: mrr_at_1000
      value: 41.048
    - type: mrr_at_3
      value: 36.718
    - type: mrr_at_5
      value: 38.841
    - type: ndcg_at_1
      value: 27.839000000000002
    - type: ndcg_at_10
      value: 44.604
    - type: ndcg_at_100
      value: 49.51
    - type: ndcg_at_1000
      value: 50.841
    - type: ndcg_at_3
      value: 37.223
    - type: ndcg_at_5
      value: 41.073
    - type: precision_at_1
      value: 27.839000000000002
    - type: precision_at_10
      value: 7.5
    - type: precision_at_100
      value: 1.03
    - type: precision_at_1000
      value: 0.116
    - type: precision_at_3
      value: 17.005
    - type: precision_at_5
      value: 12.399000000000001
    - type: recall_at_1
      value: 24.674
    - type: recall_at_10
      value: 63.32299999999999
    - type: recall_at_100
      value: 85.088
    - type: recall_at_1000
      value: 95.143
    - type: recall_at_3
      value: 44.157999999999994
    - type: recall_at_5
      value: 53.054
  - task:
      type: Classification
    dataset:
      type: laugustyniak/abusive-clauses-pl
      name: MTEB PAC
      config: default
      split: test
      revision: None
    metrics:
    - type: accuracy
      value: 64.5033304373009
    - type: ap
      value: 75.81507275237081
    - type: f1
      value: 62.24617820785985
  - task:
      type: PairClassification
    dataset:
      type: PL-MTEB/ppc-pairclassification
      name: MTEB PPC
      config: default
      split: test
      revision: None
    metrics:
    - type: cos_sim_accuracy
      value: 85.39999999999999
    - type: cos_sim_ap
      value: 91.75881977787009
    - type: cos_sim_f1
      value: 87.79264214046823
    - type: cos_sim_precision
      value: 88.68243243243244
    - type: cos_sim_recall
      value: 86.9205298013245
    - type: dot_accuracy
      value: 71.0
    - type: dot_ap
      value: 82.97829049033108
    - type: dot_f1
      value: 78.77055039313797
    - type: dot_precision
      value: 69.30817610062893
    - type: dot_recall
      value: 91.22516556291392
    - type: euclidean_accuracy
      value: 85.2
    - type: euclidean_ap
      value: 91.85245521151309
    - type: euclidean_f1
      value: 87.64607679465777
    - type: euclidean_precision
      value: 88.38383838383838
    - type: euclidean_recall
      value: 86.9205298013245
    - type: manhattan_accuracy
      value: 85.39999999999999
    - type: manhattan_ap
      value: 91.85497100160649
    - type: manhattan_f1
      value: 87.77219430485762
    - type: manhattan_precision
      value: 88.8135593220339
    - type: manhattan_recall
      value: 86.75496688741721
    - type: max_accuracy
      value: 85.39999999999999
    - type: max_ap
      value: 91.85497100160649
    - type: max_f1
      value: 87.79264214046823
  - task:
      type: PairClassification
    dataset:
      type: PL-MTEB/psc-pairclassification
      name: MTEB PSC
      config: default
      split: test
      revision: None
    metrics:
    - type: cos_sim_accuracy
      value: 97.58812615955473
    - type: cos_sim_ap
      value: 99.14945370088302
    - type: cos_sim_f1
      value: 96.06060606060606
    - type: cos_sim_precision
      value: 95.48192771084338
    - type: cos_sim_recall
      value: 96.64634146341463
    - type: dot_accuracy
      value: 95.17625231910947
    - type: dot_ap
      value: 97.05592933601112
    - type: dot_f1
      value: 92.14501510574019
    - type: dot_precision
      value: 91.31736526946108
    - type: dot_recall
      value: 92.98780487804879
    - type: euclidean_accuracy
      value: 97.6808905380334
    - type: euclidean_ap
      value: 99.18538119402824
    - type: euclidean_f1
      value: 96.20637329286798
    - type: euclidean_precision
      value: 95.77039274924472
    - type: euclidean_recall
      value: 96.64634146341463
    - type: manhattan_accuracy
      value: 97.58812615955473
    - type: manhattan_ap
      value: 99.17870990853292
    - type: manhattan_f1
      value: 96.02446483180427
    - type: manhattan_precision
      value: 96.31901840490798
    - type: manhattan_recall
      value: 95.73170731707317
    - type: max_accuracy
      value: 97.6808905380334
    - type: max_ap
      value: 99.18538119402824
    - type: max_f1
      value: 96.20637329286798
  - task:
      type: Classification
    dataset:
      type: PL-MTEB/polemo2_in
      name: MTEB PolEmo2.0-IN
      config: default
      split: test
      revision: None
    metrics:
    - type: accuracy
      value: 68.69806094182825
    - type: f1
      value: 68.0619984307764
  - task:
      type: Classification
    dataset:
      type: PL-MTEB/polemo2_out
      name: MTEB PolEmo2.0-OUT
      config: default
      split: test
      revision: None
    metrics:
    - type: accuracy
      value: 35.80971659919028
    - type: f1
      value: 31.13081621324864
  - task:
      type: Retrieval
    dataset:
      type: quora-pl
      name: MTEB Quora-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 66.149
    - type: map_at_10
      value: 80.133
    - type: map_at_100
      value: 80.845
    - type: map_at_1000
      value: 80.866
    - type: map_at_3
      value: 76.983
    - type: map_at_5
      value: 78.938
    - type: mrr_at_1
      value: 76.09
    - type: mrr_at_10
      value: 83.25099999999999
    - type: mrr_at_100
      value: 83.422
    - type: mrr_at_1000
      value: 83.42500000000001
    - type: mrr_at_3
      value: 82.02199999999999
    - type: mrr_at_5
      value: 82.831
    - type: ndcg_at_1
      value: 76.14999999999999
    - type: ndcg_at_10
      value: 84.438
    - type: ndcg_at_100
      value: 86.048
    - type: ndcg_at_1000
      value: 86.226
    - type: ndcg_at_3
      value: 80.97999999999999
    - type: ndcg_at_5
      value: 82.856
    - type: precision_at_1
      value: 76.14999999999999
    - type: precision_at_10
      value: 12.985
    - type: precision_at_100
      value: 1.513
    - type: precision_at_1000
      value: 0.156
    - type: precision_at_3
      value: 35.563
    - type: precision_at_5
      value: 23.586
    - type: recall_at_1
      value: 66.149
    - type: recall_at_10
      value: 93.195
    - type: recall_at_100
      value: 98.924
    - type: recall_at_1000
      value: 99.885
    - type: recall_at_3
      value: 83.439
    - type: recall_at_5
      value: 88.575
  - task:
      type: Retrieval
    dataset:
      type: scidocs-pl
      name: MTEB SCIDOCS-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 3.688
    - type: map_at_10
      value: 10.23
    - type: map_at_100
      value: 12.077
    - type: map_at_1000
      value: 12.382
    - type: map_at_3
      value: 7.149
    - type: map_at_5
      value: 8.689
    - type: mrr_at_1
      value: 18.2
    - type: mrr_at_10
      value: 28.816999999999997
    - type: mrr_at_100
      value: 29.982
    - type: mrr_at_1000
      value: 30.058
    - type: mrr_at_3
      value: 25.983
    - type: mrr_at_5
      value: 27.418
    - type: ndcg_at_1
      value: 18.2
    - type: ndcg_at_10
      value: 17.352999999999998
    - type: ndcg_at_100
      value: 24.859
    - type: ndcg_at_1000
      value: 30.535
    - type: ndcg_at_3
      value: 16.17
    - type: ndcg_at_5
      value: 14.235000000000001
    - type: precision_at_1
      value: 18.2
    - type: precision_at_10
      value: 9.19
    - type: precision_at_100
      value: 2.01
    - type: precision_at_1000
      value: 0.338
    - type: precision_at_3
      value: 15.5
    - type: precision_at_5
      value: 12.78
    - type: recall_at_1
      value: 3.688
    - type: recall_at_10
      value: 18.632
    - type: recall_at_100
      value: 40.822
    - type: recall_at_1000
      value: 68.552
    - type: recall_at_3
      value: 9.423
    - type: recall_at_5
      value: 12.943
  - task:
      type: PairClassification
    dataset:
      type: PL-MTEB/sicke-pl-pairclassification
      name: MTEB SICK-E-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: cos_sim_accuracy
      value: 83.12270688952303
    - type: cos_sim_ap
      value: 76.4528312253856
    - type: cos_sim_f1
      value: 68.69627507163324
    - type: cos_sim_precision
      value: 69.0922190201729
    - type: cos_sim_recall
      value: 68.30484330484332
    - type: dot_accuracy
      value: 79.20913167549939
    - type: dot_ap
      value: 65.03147071986633
    - type: dot_f1
      value: 62.812160694896846
    - type: dot_precision
      value: 50.74561403508772
    - type: dot_recall
      value: 82.4074074074074
    - type: euclidean_accuracy
      value: 83.16347329800244
    - type: euclidean_ap
      value: 76.49405838298205
    - type: euclidean_f1
      value: 68.66738120757414
    - type: euclidean_precision
      value: 68.88888888888889
    - type: euclidean_recall
      value: 68.44729344729345
    - type: manhattan_accuracy
      value: 83.16347329800244
    - type: manhattan_ap
      value: 76.5080551733795
    - type: manhattan_f1
      value: 68.73883529832084
    - type: manhattan_precision
      value: 68.9605734767025
    - type: manhattan_recall
      value: 68.51851851851852
    - type: max_accuracy
      value: 83.16347329800244
    - type: max_ap
      value: 76.5080551733795
    - type: max_f1
      value: 68.73883529832084
  - task:
      type: STS
    dataset:
      type: PL-MTEB/sickr-pl-sts
      name: MTEB SICK-R-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: cos_sim_pearson
      value: 82.60225159739653
    - type: cos_sim_spearman
      value: 76.76667220288542
    - type: euclidean_pearson
      value: 80.16302518898615
    - type: euclidean_spearman
      value: 76.76131897866455
    - type: manhattan_pearson
      value: 80.11881021613914
    - type: manhattan_spearman
      value: 76.74246419368048
  - task:
      type: STS
    dataset:
      type: mteb/sts22-crosslingual-sts
      name: MTEB STS22 (pl)
      config: pl
      split: test
      revision: 6d1ba47164174a496b7fa5d3569dae26a6813b80
    metrics:
    - type: cos_sim_pearson
      value: 38.2744776092718
    - type: cos_sim_spearman
      value: 40.35664941442517
    - type: euclidean_pearson
      value: 29.148502128336585
    - type: euclidean_spearman
      value: 40.45531563224982
    - type: manhattan_pearson
      value: 29.124177399433098
    - type: manhattan_spearman
      value: 40.2801387844354
  - task:
      type: Retrieval
    dataset:
      type: scifact-pl
      name: MTEB SciFact-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 52.994
    - type: map_at_10
      value: 63.612
    - type: map_at_100
      value: 64.294
    - type: map_at_1000
      value: 64.325
    - type: map_at_3
      value: 61.341
    - type: map_at_5
      value: 62.366
    - type: mrr_at_1
      value: 56.667
    - type: mrr_at_10
      value: 65.333
    - type: mrr_at_100
      value: 65.89399999999999
    - type: mrr_at_1000
      value: 65.91900000000001
    - type: mrr_at_3
      value: 63.666999999999994
    - type: mrr_at_5
      value: 64.36699999999999
    - type: ndcg_at_1
      value: 56.333
    - type: ndcg_at_10
      value: 68.292
    - type: ndcg_at_100
      value: 71.136
    - type: ndcg_at_1000
      value: 71.90100000000001
    - type: ndcg_at_3
      value: 64.387
    - type: ndcg_at_5
      value: 65.546
    - type: precision_at_1
      value: 56.333
    - type: precision_at_10
      value: 9.133
    - type: precision_at_100
      value: 1.0630000000000002
    - type: precision_at_1000
      value: 0.11299999999999999
    - type: precision_at_3
      value: 25.556
    - type: precision_at_5
      value: 16.267
    - type: recall_at_1
      value: 52.994
    - type: recall_at_10
      value: 81.178
    - type: recall_at_100
      value: 93.767
    - type: recall_at_1000
      value: 99.667
    - type: recall_at_3
      value: 69.906
    - type: recall_at_5
      value: 73.18299999999999
  - task:
      type: Retrieval
    dataset:
      type: trec-covid-pl
      name: MTEB TRECCOVID-PL
      config: default
      split: test
      revision: None
    metrics:
    - type: map_at_1
      value: 0.231
    - type: map_at_10
      value: 1.822
    - type: map_at_100
      value: 10.134
    - type: map_at_1000
      value: 24.859
    - type: map_at_3
      value: 0.615
    - type: map_at_5
      value: 0.9939999999999999
    - type: mrr_at_1
      value: 84.0
    - type: mrr_at_10
      value: 90.4
    - type: mrr_at_100
      value: 90.4
    - type: mrr_at_1000
      value: 90.4
    - type: mrr_at_3
      value: 89.0
    - type: mrr_at_5
      value: 90.4
    - type: ndcg_at_1
      value: 81.0
    - type: ndcg_at_10
      value: 73.333
    - type: ndcg_at_100
      value: 55.35099999999999
    - type: ndcg_at_1000
      value: 49.875
    - type: ndcg_at_3
      value: 76.866
    - type: ndcg_at_5
      value: 75.472
    - type: precision_at_1
      value: 86.0
    - type: precision_at_10
      value: 78.2
    - type: precision_at_100
      value: 57.18
    - type: precision_at_1000
      value: 22.332
    - type: precision_at_3
      value: 82.0
    - type: precision_at_5
      value: 81.2
    - type: recall_at_1
      value: 0.231
    - type: recall_at_10
      value: 2.056
    - type: recall_at_100
      value: 13.468
    - type: recall_at_1000
      value: 47.038999999999994
    - type: recall_at_3
      value: 0.6479999999999999
    - type: recall_at_5
      value: 1.088
language: pl
license: apache-2.0
widget:
- source_sentence: "query: Jak dożyć 100 lat?"
  sentences:
    - "passage: Trzeba zdrowo się odżywiać i uprawiać sport."
    - "passage: Trzeba pić alkohol, imprezować i jeździć szybkimi autami."
    - "passage: Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."

---

<h1 align="center">MMLW-e5-base</h1>

MMLW (muszę mieć lepszą wiadomość) are neural text encoders for Polish.
This is a distilled model that can be used to generate embeddings applicable to many tasks such as semantic similarity, clustering, information retrieval. The model can also serve as a base for further fine-tuning. 
It transforms texts to 768 dimensional vectors. 
The model was initialized with multilingual E5 checkpoint, and then trained with [multilingual knowledge distillation method](https://aclanthology.org/2020.emnlp-main.365/) on a diverse corpus of 60 million Polish-English text pairs. We utilised [English FlagEmbeddings (BGE)](https://huggingface.co/BAAI/bge-base-en) as teacher models for distillation. 

## Usage (Sentence-Transformers)

⚠️ Our embedding models require the use of specific prefixes and suffixes when encoding texts. For this model, queries should be prefixed with **"query: "** and passages with **"passage: "** ⚠️

You can use the model like this with [sentence-transformers](https://www.SBERT.net):

```python
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim

query_prefix = "query: "
answer_prefix = "passage: "
queries = [query_prefix + "Jak dożyć 100 lat?"]
answers = [
    answer_prefix + "Trzeba zdrowo się odżywiać i uprawiać sport.",
    answer_prefix + "Trzeba pić alkohol, imprezować i jeździć szybkimi autami.",
    answer_prefix + "Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
]
model = SentenceTransformer("sdadas/mmlw-e5-base")
queries_emb = model.encode(queries, convert_to_tensor=True, show_progress_bar=False)
answers_emb = model.encode(answers, convert_to_tensor=True, show_progress_bar=False)

best_answer = cos_sim(queries_emb, answers_emb).argmax().item()
print(answers[best_answer])
# Trzeba zdrowo się odżywiać i uprawiać sport.
```

## Evaluation Results

- The model achieves an **Average Score** of **59.71** on the Polish Massive Text Embedding Benchmark (MTEB). See [MTEB Leaderboard](https://huggingface.co/spaces/mteb/leaderboard) for detailed results.
- The model achieves **NDCG@10** of **53.56** on the Polish Information Retrieval Benchmark. See [PIRB Leaderboard](https://huggingface.co/spaces/sdadas/pirb) for detailed results.

## Acknowledgements
This model was trained with the A100 GPU cluster support delivered by the Gdansk University of Technology within the TASK center initiative.