--- tags: - mteb - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:2560000 - loss:MultipleNegativesRankingLoss model-index: - name: XLM-RoBERTa-base-MSMARCO-WebFAQ results: - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (ar) config: ar split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 25.374000000000002 - type: map_at_10 value: 40.863 - type: map_at_100 value: 42.667 - type: map_at_1000 value: 42.754999999999995 - type: map_at_20 value: 41.948 - type: map_at_3 value: 35.512 - type: map_at_5 value: 38.385999999999996 - type: mrr_at_1 value: 38.6 - type: mrr_at_10 value: 50.489 - type: mrr_at_100 value: 51.229 - type: mrr_at_1000 value: 51.248000000000005 - type: mrr_at_20 value: 50.980000000000004 - type: mrr_at_3 value: 46.933 - type: mrr_at_5 value: 48.983 - type: ndcg_at_1 value: 38.6 - type: ndcg_at_10 value: 49.257 - type: ndcg_at_100 value: 55.611999999999995 - type: ndcg_at_1000 value: 56.946 - type: ndcg_at_20 value: 52.10399999999999 - type: ndcg_at_3 value: 41.501 - type: ndcg_at_5 value: 44.729 - type: precision_at_1 value: 38.6 - type: precision_at_10 value: 11.540000000000001 - type: precision_at_100 value: 1.702 - type: precision_at_1000 value: 0.189 - type: precision_at_20 value: 6.795 - type: precision_at_3 value: 24.3 - type: precision_at_5 value: 18.16 - type: recall_at_1 value: 25.374000000000002 - type: recall_at_10 value: 63.474 - type: recall_at_100 value: 87.902 - type: recall_at_1000 value: 96.301 - type: recall_at_20 value: 72.538 - type: recall_at_3 value: 43.144 - type: recall_at_5 value: 51.548 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (arabic) config: arabic split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 26.056 - type: map_at_10 value: 37.684 - type: map_at_100 value: 38.616 - type: map_at_1000 value: 38.66 - type: map_at_20 value: 38.281 - type: map_at_3 value: 34.348 - type: map_at_5 value: 36.181999999999995 - type: mrr_at_1 value: 27.845 - type: mrr_at_10 value: 39.275 - type: mrr_at_100 value: 40.083 - type: mrr_at_1000 value: 40.115 - type: mrr_at_20 value: 39.822 - type: mrr_at_3 value: 36.124 - type: mrr_at_5 value: 37.885999999999996 - type: ndcg_at_1 value: 27.845 - type: ndcg_at_10 value: 44.153 - type: ndcg_at_100 value: 48.291000000000004 - type: ndcg_at_1000 value: 49.443 - type: ndcg_at_20 value: 46.188 - type: ndcg_at_3 value: 37.448 - type: ndcg_at_5 value: 40.685 - type: precision_at_1 value: 27.845 - type: precision_at_10 value: 6.938 - type: precision_at_100 value: 0.9209999999999999 - type: precision_at_1000 value: 0.10300000000000001 - type: precision_at_20 value: 3.936 - type: precision_at_3 value: 16.158 - type: precision_at_5 value: 11.452 - type: recall_at_1 value: 26.056 - type: recall_at_10 value: 62.627 - type: recall_at_100 value: 80.759 - type: recall_at_1000 value: 89.547 - type: recall_at_20 value: 70.21300000000001 - type: recall_at_3 value: 44.681 - type: recall_at_5 value: 52.344 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (bn) config: bn split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 23.083000000000002 - type: map_at_10 value: 37.285000000000004 - type: map_at_100 value: 38.834 - type: map_at_1000 value: 38.948 - type: map_at_20 value: 38.195 - type: map_at_3 value: 31.893 - type: map_at_5 value: 34.755 - type: mrr_at_1 value: 36.496 - type: mrr_at_10 value: 48.486000000000004 - type: mrr_at_100 value: 49.248 - type: mrr_at_1000 value: 49.28 - type: mrr_at_20 value: 48.998000000000005 - type: mrr_at_3 value: 44.85 - type: mrr_at_5 value: 47.028 - type: ndcg_at_1 value: 36.496 - type: ndcg_at_10 value: 45.96 - type: ndcg_at_100 value: 51.64 - type: ndcg_at_1000 value: 53.542 - type: ndcg_at_20 value: 48.543 - type: ndcg_at_3 value: 37.964999999999996 - type: ndcg_at_5 value: 41.44 - type: precision_at_1 value: 36.496 - type: precision_at_10 value: 11.29 - type: precision_at_100 value: 1.6199999999999999 - type: precision_at_1000 value: 0.191 - type: precision_at_20 value: 6.582000000000001 - type: precision_at_3 value: 22.871 - type: precision_at_5 value: 17.324 - type: recall_at_1 value: 23.083000000000002 - type: recall_at_10 value: 59.414 - type: recall_at_100 value: 81.08 - type: recall_at_1000 value: 92.793 - type: recall_at_20 value: 67.634 - type: recall_at_3 value: 39.001000000000005 - type: recall_at_5 value: 47.612 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (bengali) config: bengali split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 23.874000000000002 - type: map_at_10 value: 38.235 - type: map_at_100 value: 39.428000000000004 - type: map_at_1000 value: 39.458 - type: map_at_20 value: 39.119 - type: map_at_3 value: 34.233999999999995 - type: map_at_5 value: 36.577 - type: mrr_at_1 value: 28.829 - type: mrr_at_10 value: 41.032999999999994 - type: mrr_at_100 value: 42.077999999999996 - type: mrr_at_1000 value: 42.104 - type: mrr_at_20 value: 41.855 - type: mrr_at_3 value: 37.688 - type: mrr_at_5 value: 39.985 - type: ndcg_at_1 value: 28.829 - type: ndcg_at_10 value: 45.214999999999996 - type: ndcg_at_100 value: 49.986999999999995 - type: ndcg_at_1000 value: 50.67700000000001 - type: ndcg_at_20 value: 48.291000000000004 - type: ndcg_at_3 value: 37.822 - type: ndcg_at_5 value: 41.9 - type: precision_at_1 value: 28.829 - type: precision_at_10 value: 7.477 - type: precision_at_100 value: 0.9820000000000001 - type: precision_at_1000 value: 0.105 - type: precision_at_20 value: 4.414 - type: precision_at_3 value: 17.416999999999998 - type: precision_at_5 value: 12.613 - type: recall_at_1 value: 23.874000000000002 - type: recall_at_10 value: 63.964 - type: recall_at_100 value: 83.784 - type: recall_at_1000 value: 88.739 - type: recall_at_20 value: 75.676 - type: recall_at_3 value: 45.045 - type: recall_at_5 value: 54.505 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (de) config: de split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 12.524 - type: map_at_10 value: 25.932 - type: map_at_100 value: 29.215999999999998 - type: map_at_1000 value: 29.342000000000002 - type: map_at_20 value: 27.515 - type: map_at_3 value: 20.196 - type: map_at_5 value: 23.311999999999998 - type: mrr_at_1 value: 28.525 - type: mrr_at_10 value: 41.854 - type: mrr_at_100 value: 42.977 - type: mrr_at_1000 value: 42.998 - type: mrr_at_20 value: 42.628 - type: mrr_at_3 value: 38.251000000000005 - type: mrr_at_5 value: 40.514 - type: ndcg_at_1 value: 28.525 - type: ndcg_at_10 value: 35.831 - type: ndcg_at_100 value: 47.410000000000004 - type: ndcg_at_1000 value: 48.931000000000004 - type: ndcg_at_20 value: 40.021 - type: ndcg_at_3 value: 28.614 - type: ndcg_at_5 value: 31.427 - type: precision_at_1 value: 28.525 - type: precision_at_10 value: 11.934000000000001 - type: precision_at_100 value: 2.37 - type: precision_at_1000 value: 0.261 - type: precision_at_20 value: 7.704999999999999 - type: precision_at_3 value: 21.421 - type: precision_at_5 value: 17.574 - type: recall_at_1 value: 12.524 - type: recall_at_10 value: 47.558 - type: recall_at_100 value: 90.068 - type: recall_at_1000 value: 98.68900000000001 - type: recall_at_20 value: 59.943999999999996 - type: recall_at_3 value: 26.889999999999997 - type: recall_at_5 value: 36.152 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (en) config: en split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 14.431 - type: map_at_10 value: 27.194000000000003 - type: map_at_100 value: 30.575000000000003 - type: map_at_1000 value: 30.75 - type: map_at_20 value: 28.938999999999997 - type: map_at_3 value: 21.69 - type: map_at_5 value: 24.657999999999998 - type: mrr_at_1 value: 31.163999999999998 - type: mrr_at_10 value: 43.417 - type: mrr_at_100 value: 44.48 - type: mrr_at_1000 value: 44.499 - type: mrr_at_20 value: 44.112 - type: mrr_at_3 value: 39.57 - type: mrr_at_5 value: 41.692 - type: ndcg_at_1 value: 31.163999999999998 - type: ndcg_at_10 value: 36.925000000000004 - type: ndcg_at_100 value: 48.15 - type: ndcg_at_1000 value: 50.131 - type: ndcg_at_20 value: 41.308 - type: ndcg_at_3 value: 30.291 - type: ndcg_at_5 value: 32.731 - type: precision_at_1 value: 31.163999999999998 - type: precision_at_10 value: 12.065 - type: precision_at_100 value: 2.469 - type: precision_at_1000 value: 0.28500000000000003 - type: precision_at_20 value: 7.997 - type: precision_at_3 value: 21.443 - type: precision_at_5 value: 17.447 - type: recall_at_1 value: 14.431 - type: recall_at_10 value: 48.275 - type: recall_at_100 value: 87.19200000000001 - type: recall_at_1000 value: 97.79299999999999 - type: recall_at_20 value: 60.870000000000005 - type: recall_at_3 value: 27.679 - type: recall_at_5 value: 36.479 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (english) config: english split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 13.642000000000001 - type: map_at_10 value: 22.744 - type: map_at_100 value: 24.083 - type: map_at_1000 value: 24.157 - type: map_at_20 value: 23.592 - type: map_at_3 value: 19.134999999999998 - type: map_at_5 value: 21.448999999999998 - type: mrr_at_1 value: 16.667 - type: mrr_at_10 value: 25.77 - type: mrr_at_100 value: 26.878 - type: mrr_at_1000 value: 26.938000000000002 - type: mrr_at_20 value: 26.503 - type: mrr_at_3 value: 22.512999999999998 - type: mrr_at_5 value: 24.516 - type: ndcg_at_1 value: 16.667 - type: ndcg_at_10 value: 28.652 - type: ndcg_at_100 value: 34.716 - type: ndcg_at_1000 value: 36.649 - type: ndcg_at_20 value: 31.508000000000003 - type: ndcg_at_3 value: 21.748 - type: ndcg_at_5 value: 25.647 - type: precision_at_1 value: 16.667 - type: precision_at_10 value: 5.3629999999999995 - type: precision_at_100 value: 0.886 - type: precision_at_1000 value: 0.107 - type: precision_at_20 value: 3.36 - type: precision_at_3 value: 10.529 - type: precision_at_5 value: 8.522 - type: recall_at_1 value: 13.642000000000001 - type: recall_at_10 value: 43.884 - type: recall_at_100 value: 70.744 - type: recall_at_1000 value: 85.372 - type: recall_at_20 value: 54.48 - type: recall_at_3 value: 26.142 - type: recall_at_5 value: 35.17 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (fa) config: fa split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 20.752000000000002 - type: map_at_10 value: 34.931 - type: map_at_100 value: 37.132 - type: map_at_1000 value: 37.230999999999995 - type: map_at_20 value: 36.144 - type: map_at_3 value: 29.692 - type: map_at_5 value: 32.536 - type: mrr_at_1 value: 32.437 - type: mrr_at_10 value: 44.41 - type: mrr_at_100 value: 45.324 - type: mrr_at_1000 value: 45.348 - type: mrr_at_20 value: 44.945 - type: mrr_at_3 value: 41.482 - type: mrr_at_5 value: 43.183 - type: ndcg_at_1 value: 32.437 - type: ndcg_at_10 value: 43.018 - type: ndcg_at_100 value: 50.805 - type: ndcg_at_1000 value: 52.245 - type: ndcg_at_20 value: 46.215 - type: ndcg_at_3 value: 36.269 - type: ndcg_at_5 value: 39.101 - type: precision_at_1 value: 32.437 - type: precision_at_10 value: 11.013 - type: precision_at_100 value: 1.799 - type: precision_at_1000 value: 0.201 - type: precision_at_20 value: 6.7250000000000005 - type: precision_at_3 value: 22.416 - type: precision_at_5 value: 17.152 - type: recall_at_1 value: 20.752000000000002 - type: recall_at_10 value: 56.150999999999996 - type: recall_at_100 value: 85.735 - type: recall_at_1000 value: 94.599 - type: recall_at_20 value: 66.237 - type: recall_at_3 value: 37.551 - type: recall_at_5 value: 45.629 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (fi) config: fi split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 28.829 - type: map_at_10 value: 46.749 - type: map_at_100 value: 48.457 - type: map_at_1000 value: 48.516 - type: map_at_20 value: 47.798 - type: map_at_3 value: 40.449 - type: map_at_5 value: 44.3 - type: mrr_at_1 value: 46.5 - type: mrr_at_10 value: 58.619 - type: mrr_at_100 value: 59.294999999999995 - type: mrr_at_1000 value: 59.307 - type: mrr_at_20 value: 59.071 - type: mrr_at_3 value: 55.75 - type: mrr_at_5 value: 57.605 - type: ndcg_at_1 value: 46.5 - type: ndcg_at_10 value: 55.933 - type: ndcg_at_100 value: 61.732 - type: ndcg_at_1000 value: 62.651 - type: ndcg_at_20 value: 58.679 - type: ndcg_at_3 value: 46.866 - type: ndcg_at_5 value: 51.625 - type: precision_at_1 value: 46.5 - type: precision_at_10 value: 12.920000000000002 - type: precision_at_100 value: 1.745 - type: precision_at_1000 value: 0.187 - type: precision_at_20 value: 7.37 - type: precision_at_3 value: 28.166999999999998 - type: precision_at_5 value: 21.2 - type: recall_at_1 value: 28.829 - type: recall_at_10 value: 70.075 - type: recall_at_100 value: 92.098 - type: recall_at_1000 value: 97.813 - type: recall_at_20 value: 78.975 - type: recall_at_3 value: 48.635 - type: recall_at_5 value: 59.202999999999996 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (finnish) config: finnish split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 20.654 - type: map_at_10 value: 32.103 - type: map_at_100 value: 33.129 - type: map_at_1000 value: 33.184999999999995 - type: map_at_20 value: 32.739000000000004 - type: map_at_3 value: 28.338 - type: map_at_5 value: 30.564000000000004 - type: mrr_at_1 value: 22.567999999999998 - type: mrr_at_10 value: 33.949 - type: mrr_at_100 value: 34.804 - type: mrr_at_1000 value: 34.849999999999994 - type: mrr_at_20 value: 34.496 - type: mrr_at_3 value: 30.581999999999997 - type: mrr_at_5 value: 32.5 - type: ndcg_at_1 value: 22.567999999999998 - type: ndcg_at_10 value: 38.681 - type: ndcg_at_100 value: 43.367 - type: ndcg_at_1000 value: 44.836999999999996 - type: ndcg_at_20 value: 40.815 - type: ndcg_at_3 value: 31.328 - type: ndcg_at_5 value: 35.083 - type: precision_at_1 value: 22.567999999999998 - type: precision_at_10 value: 6.483 - type: precision_at_100 value: 0.906 - type: precision_at_1000 value: 0.104 - type: precision_at_20 value: 3.7319999999999998 - type: precision_at_3 value: 14.035 - type: precision_at_5 value: 10.51 - type: recall_at_1 value: 20.654 - type: recall_at_10 value: 57.297 - type: recall_at_100 value: 78.363 - type: recall_at_1000 value: 89.673 - type: recall_at_20 value: 65.28399999999999 - type: recall_at_3 value: 37.799 - type: recall_at_5 value: 46.518 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (fr) config: fr split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 15.484 - type: map_at_10 value: 29.692 - type: map_at_100 value: 32.604 - type: map_at_1000 value: 32.668 - type: map_at_20 value: 31.397000000000002 - type: map_at_3 value: 23.469 - type: map_at_5 value: 26.454 - type: mrr_at_1 value: 28.863 - type: mrr_at_10 value: 41.03 - type: mrr_at_100 value: 42.134 - type: mrr_at_1000 value: 42.149 - type: mrr_at_20 value: 41.730000000000004 - type: mrr_at_3 value: 36.394999999999996 - type: mrr_at_5 value: 38.814 - type: ndcg_at_1 value: 28.863 - type: ndcg_at_10 value: 39.523 - type: ndcg_at_100 value: 49.496 - type: ndcg_at_1000 value: 50.375 - type: ndcg_at_20 value: 43.923 - type: ndcg_at_3 value: 29.309 - type: ndcg_at_5 value: 33.077 - type: precision_at_1 value: 28.863 - type: precision_at_10 value: 11.254 - type: precision_at_100 value: 1.997 - type: precision_at_1000 value: 0.212 - type: precision_at_20 value: 7.172000000000001 - type: precision_at_3 value: 19.631 - type: precision_at_5 value: 15.568999999999999 - type: recall_at_1 value: 15.484 - type: recall_at_10 value: 56.279999999999994 - type: recall_at_100 value: 93.714 - type: recall_at_1000 value: 99.125 - type: recall_at_20 value: 70.215 - type: recall_at_3 value: 29.688 - type: recall_at_5 value: 39.329 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (hi) config: hi split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 16.442999999999998 - type: map_at_10 value: 29.224 - type: map_at_100 value: 31.386999999999997 - type: map_at_1000 value: 31.529 - type: map_at_20 value: 30.54 - type: map_at_3 value: 24.521 - type: map_at_5 value: 26.979 - type: mrr_at_1 value: 33.428999999999995 - type: mrr_at_10 value: 42.951 - type: mrr_at_100 value: 43.832 - type: mrr_at_1000 value: 43.891000000000005 - type: mrr_at_20 value: 43.494 - type: mrr_at_3 value: 40.048 - type: mrr_at_5 value: 41.819 - type: ndcg_at_1 value: 33.428999999999995 - type: ndcg_at_10 value: 37.462 - type: ndcg_at_100 value: 45.123000000000005 - type: ndcg_at_1000 value: 47.805 - type: ndcg_at_20 value: 40.739999999999995 - type: ndcg_at_3 value: 31.89 - type: ndcg_at_5 value: 33.934999999999995 - type: precision_at_1 value: 33.428999999999995 - type: precision_at_10 value: 10.343 - type: precision_at_100 value: 1.714 - type: precision_at_1000 value: 0.20500000000000002 - type: precision_at_20 value: 6.4 - type: precision_at_3 value: 20.952 - type: precision_at_5 value: 15.886 - type: recall_at_1 value: 16.442999999999998 - type: recall_at_10 value: 47.987 - type: recall_at_100 value: 77.628 - type: recall_at_1000 value: 95.673 - type: recall_at_20 value: 58.199999999999996 - type: recall_at_3 value: 30.904999999999998 - type: recall_at_5 value: 38.705 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (id) config: id split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 14.551 - type: map_at_10 value: 27.144000000000002 - type: map_at_100 value: 29.918 - type: map_at_1000 value: 30.14 - type: map_at_20 value: 28.665000000000003 - type: map_at_3 value: 22.004 - type: map_at_5 value: 24.851 - type: mrr_at_1 value: 33.542 - type: mrr_at_10 value: 45.576 - type: mrr_at_100 value: 46.402 - type: mrr_at_1000 value: 46.432 - type: mrr_at_20 value: 46.082 - type: mrr_at_3 value: 42.483 - type: mrr_at_5 value: 44.399 - type: ndcg_at_1 value: 33.542 - type: ndcg_at_10 value: 36.287000000000006 - type: ndcg_at_100 value: 45.253 - type: ndcg_at_1000 value: 48.33 - type: ndcg_at_20 value: 39.855000000000004 - type: ndcg_at_3 value: 31.69 - type: ndcg_at_5 value: 33.332 - type: precision_at_1 value: 33.542 - type: precision_at_10 value: 12.333 - type: precision_at_100 value: 2.322 - type: precision_at_1000 value: 0.292 - type: precision_at_20 value: 7.875 - type: precision_at_3 value: 23.125 - type: precision_at_5 value: 18.583 - type: recall_at_1 value: 14.551 - type: recall_at_10 value: 43.636 - type: recall_at_100 value: 73.603 - type: recall_at_1000 value: 90.596 - type: recall_at_20 value: 53.559 - type: recall_at_3 value: 27.383999999999997 - type: recall_at_5 value: 34.997 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (indonesian) config: indonesian split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 28.669 - type: map_at_10 value: 41.611 - type: map_at_100 value: 42.522999999999996 - type: map_at_1000 value: 42.552 - type: map_at_20 value: 42.211 - type: map_at_3 value: 37.719 - type: map_at_5 value: 40.182 - type: mrr_at_1 value: 31.846000000000004 - type: mrr_at_10 value: 43.695 - type: mrr_at_100 value: 44.45 - type: mrr_at_1000 value: 44.474000000000004 - type: mrr_at_20 value: 44.214999999999996 - type: mrr_at_3 value: 40.31 - type: mrr_at_5 value: 42.469 - type: ndcg_at_1 value: 31.846000000000004 - type: ndcg_at_10 value: 48.416 - type: ndcg_at_100 value: 52.464999999999996 - type: ndcg_at_1000 value: 53.234 - type: ndcg_at_20 value: 50.468999999999994 - type: ndcg_at_3 value: 40.973 - type: ndcg_at_5 value: 45.163 - type: precision_at_1 value: 31.846000000000004 - type: precision_at_10 value: 7.768 - type: precision_at_100 value: 0.992 - type: precision_at_1000 value: 0.106 - type: precision_at_20 value: 4.343 - type: precision_at_3 value: 18.134 - type: precision_at_5 value: 13.221 - type: recall_at_1 value: 28.669 - type: recall_at_10 value: 67.29 - type: recall_at_100 value: 85.324 - type: recall_at_1000 value: 91.27499999999999 - type: recall_at_20 value: 75.111 - type: recall_at_3 value: 47.869 - type: recall_at_5 value: 57.620000000000005 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (ja) config: ja split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 16.499 - type: map_at_10 value: 29.566 - type: map_at_100 value: 31.846999999999998 - type: map_at_1000 value: 31.968999999999998 - type: map_at_20 value: 30.758000000000003 - type: map_at_3 value: 24.954 - type: map_at_5 value: 27.195000000000004 - type: mrr_at_1 value: 27.093 - type: mrr_at_10 value: 40.861999999999995 - type: mrr_at_100 value: 41.926 - type: mrr_at_1000 value: 41.942 - type: mrr_at_20 value: 41.55 - type: mrr_at_3 value: 37.151 - type: mrr_at_5 value: 38.958999999999996 - type: ndcg_at_1 value: 27.093 - type: ndcg_at_10 value: 38.544 - type: ndcg_at_100 value: 47.143 - type: ndcg_at_1000 value: 48.802 - type: ndcg_at_20 value: 41.896 - type: ndcg_at_3 value: 31.249 - type: ndcg_at_5 value: 33.873999999999995 - type: precision_at_1 value: 27.093 - type: precision_at_10 value: 9.814 - type: precision_at_100 value: 1.737 - type: precision_at_1000 value: 0.2 - type: precision_at_20 value: 6.064 - type: precision_at_3 value: 19.147 - type: precision_at_5 value: 14.465 - type: recall_at_1 value: 16.499 - type: recall_at_10 value: 53.580000000000005 - type: recall_at_100 value: 86.792 - type: recall_at_1000 value: 96.61999999999999 - type: recall_at_20 value: 64.302 - type: recall_at_3 value: 33.588 - type: recall_at_5 value: 40.993 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (japanese) config: japanese split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 16.644000000000002 - type: map_at_10 value: 25.667 - type: map_at_100 value: 26.926 - type: map_at_1000 value: 27.003 - type: map_at_20 value: 26.468000000000004 - type: map_at_3 value: 22.808999999999997 - type: map_at_5 value: 24.392 - type: mrr_at_1 value: 20.278 - type: mrr_at_10 value: 28.683999999999997 - type: mrr_at_100 value: 29.785 - type: mrr_at_1000 value: 29.839 - type: mrr_at_20 value: 29.391000000000002 - type: mrr_at_3 value: 26.25 - type: mrr_at_5 value: 27.479 - type: ndcg_at_1 value: 20.278 - type: ndcg_at_10 value: 31.130000000000003 - type: ndcg_at_100 value: 36.954 - type: ndcg_at_1000 value: 38.805 - type: ndcg_at_20 value: 33.856 - type: ndcg_at_3 value: 25.590000000000003 - type: ndcg_at_5 value: 28.136 - type: precision_at_1 value: 20.278 - type: precision_at_10 value: 5.611 - type: precision_at_100 value: 0.899 - type: precision_at_1000 value: 0.109 - type: precision_at_20 value: 3.465 - type: precision_at_3 value: 12.5 - type: precision_at_5 value: 8.972 - type: recall_at_1 value: 16.644000000000002 - type: recall_at_10 value: 44.931 - type: recall_at_100 value: 70.741 - type: recall_at_1000 value: 84.282 - type: recall_at_20 value: 55.022999999999996 - type: recall_at_3 value: 30.37 - type: recall_at_5 value: 36.134 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (ko) config: ko split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 20.192 - type: map_at_10 value: 32.281 - type: map_at_100 value: 34.396 - type: map_at_1000 value: 34.604 - type: map_at_20 value: 33.495000000000005 - type: map_at_3 value: 27.489 - type: map_at_5 value: 30.022 - type: mrr_at_1 value: 34.742 - type: mrr_at_10 value: 46.117999999999995 - type: mrr_at_100 value: 47.066 - type: mrr_at_1000 value: 47.095 - type: mrr_at_20 value: 46.867 - type: mrr_at_3 value: 41.862 - type: mrr_at_5 value: 44.702999999999996 - type: ndcg_at_1 value: 34.742 - type: ndcg_at_10 value: 41.193999999999996 - type: ndcg_at_100 value: 48.691 - type: ndcg_at_1000 value: 51.364 - type: ndcg_at_20 value: 44.592999999999996 - type: ndcg_at_3 value: 35.004000000000005 - type: ndcg_at_5 value: 37.608000000000004 - type: precision_at_1 value: 34.742 - type: precision_at_10 value: 10.61 - type: precision_at_100 value: 1.8450000000000002 - type: precision_at_1000 value: 0.24 - type: precision_at_20 value: 6.5729999999999995 - type: precision_at_3 value: 20.657 - type: precision_at_5 value: 16.150000000000002 - type: recall_at_1 value: 20.192 - type: recall_at_10 value: 54.154 - type: recall_at_100 value: 80.49199999999999 - type: recall_at_1000 value: 94.61699999999999 - type: recall_at_20 value: 64.74 - type: recall_at_3 value: 34.288000000000004 - type: recall_at_5 value: 43.401 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (korean) config: korean split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 22.288 - type: map_at_10 value: 30.096 - type: map_at_100 value: 30.866 - type: map_at_1000 value: 30.939 - type: map_at_20 value: 30.496000000000002 - type: map_at_3 value: 27.672 - type: map_at_5 value: 28.866000000000003 - type: mrr_at_1 value: 23.990000000000002 - type: mrr_at_10 value: 32.017 - type: mrr_at_100 value: 32.665 - type: mrr_at_1000 value: 32.726 - type: mrr_at_20 value: 32.348 - type: mrr_at_3 value: 29.572 - type: mrr_at_5 value: 30.808000000000003 - type: ndcg_at_1 value: 23.990000000000002 - type: ndcg_at_10 value: 34.823 - type: ndcg_at_100 value: 38.625 - type: ndcg_at_1000 value: 40.760999999999996 - type: ndcg_at_20 value: 36.138 - type: ndcg_at_3 value: 29.744999999999997 - type: ndcg_at_5 value: 31.884 - type: precision_at_1 value: 23.990000000000002 - type: precision_at_10 value: 5.297000000000001 - type: precision_at_100 value: 0.753 - type: precision_at_1000 value: 0.095 - type: precision_at_20 value: 2.981 - type: precision_at_3 value: 12.272 - type: precision_at_5 value: 8.551 - type: recall_at_1 value: 22.288 - type: recall_at_10 value: 47.743 - type: recall_at_100 value: 65.202 - type: recall_at_1000 value: 82.106 - type: recall_at_20 value: 52.534000000000006 - type: recall_at_3 value: 33.927 - type: recall_at_5 value: 38.915 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (ru) config: ru split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 16.445 - type: map_at_10 value: 29.376 - type: map_at_100 value: 32.425 - type: map_at_1000 value: 32.589 - type: map_at_20 value: 30.973 - type: map_at_3 value: 24.026 - type: map_at_5 value: 26.831 - type: mrr_at_1 value: 31.8 - type: mrr_at_10 value: 44.454 - type: mrr_at_100 value: 45.534 - type: mrr_at_1000 value: 45.552 - type: mrr_at_20 value: 45.192 - type: mrr_at_3 value: 41.117 - type: mrr_at_5 value: 43.132 - type: ndcg_at_1 value: 31.8 - type: ndcg_at_10 value: 38.487 - type: ndcg_at_100 value: 48.705999999999996 - type: ndcg_at_1000 value: 50.732 - type: ndcg_at_20 value: 42.646 - type: ndcg_at_3 value: 31.922 - type: ndcg_at_5 value: 34.512 - type: precision_at_1 value: 31.8 - type: precision_at_10 value: 12.01 - type: precision_at_100 value: 2.325 - type: precision_at_1000 value: 0.27 - type: precision_at_20 value: 7.8549999999999995 - type: precision_at_3 value: 21.867 - type: precision_at_5 value: 17.46 - type: recall_at_1 value: 16.445 - type: recall_at_10 value: 48.691 - type: recall_at_100 value: 84.084 - type: recall_at_1000 value: 95.318 - type: recall_at_20 value: 60.873 - type: recall_at_3 value: 30.169 - type: recall_at_5 value: 38.3 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (russian) config: russian split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 16.516000000000002 - type: map_at_10 value: 24.461 - type: map_at_100 value: 25.499 - type: map_at_1000 value: 25.569999999999997 - type: map_at_20 value: 25.063000000000002 - type: map_at_3 value: 21.669 - type: map_at_5 value: 23.285 - type: mrr_at_1 value: 18.09 - type: mrr_at_10 value: 26.311 - type: mrr_at_100 value: 27.195999999999998 - type: mrr_at_1000 value: 27.255000000000003 - type: mrr_at_20 value: 26.823000000000004 - type: mrr_at_3 value: 23.618 - type: mrr_at_5 value: 25.156 - type: ndcg_at_1 value: 18.09 - type: ndcg_at_10 value: 29.453000000000003 - type: ndcg_at_100 value: 34.44 - type: ndcg_at_1000 value: 36.336 - type: ndcg_at_20 value: 31.482 - type: ndcg_at_3 value: 23.830000000000002 - type: ndcg_at_5 value: 26.666 - type: precision_at_1 value: 18.09 - type: precision_at_10 value: 4.925 - type: precision_at_100 value: 0.768 - type: precision_at_1000 value: 0.094 - type: precision_at_20 value: 2.9250000000000003 - type: precision_at_3 value: 10.586 - type: precision_at_5 value: 7.839 - type: recall_at_1 value: 16.516000000000002 - type: recall_at_10 value: 43.166 - type: recall_at_100 value: 66.11399999999999 - type: recall_at_1000 value: 80.771 - type: recall_at_20 value: 50.804 - type: recall_at_3 value: 28.191 - type: recall_at_5 value: 34.925 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (es) config: es split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 12.144 - type: map_at_10 value: 28.499999999999996 - type: map_at_100 value: 34.048 - type: map_at_1000 value: 34.268 - type: map_at_20 value: 31.401 - type: map_at_3 value: 20.459 - type: map_at_5 value: 24.245 - type: mrr_at_1 value: 41.975 - type: mrr_at_10 value: 54.525 - type: mrr_at_100 value: 55.359 - type: mrr_at_1000 value: 55.367 - type: mrr_at_20 value: 55.1 - type: mrr_at_3 value: 51.929 - type: mrr_at_5 value: 53.418 - type: ndcg_at_1 value: 41.975 - type: ndcg_at_10 value: 40.117000000000004 - type: ndcg_at_100 value: 54.102 - type: ndcg_at_1000 value: 56.191 - type: ndcg_at_20 value: 45.67 - type: ndcg_at_3 value: 37.505 - type: ndcg_at_5 value: 36.968 - type: precision_at_1 value: 41.975 - type: precision_at_10 value: 19.707 - type: precision_at_100 value: 4.003 - type: precision_at_1000 value: 0.44799999999999995 - type: precision_at_20 value: 13.272 - type: precision_at_3 value: 31.790000000000003 - type: precision_at_5 value: 26.667 - type: recall_at_1 value: 12.144 - type: recall_at_10 value: 44.92 - type: recall_at_100 value: 86.141 - type: recall_at_1000 value: 96.234 - type: recall_at_20 value: 58.194 - type: recall_at_3 value: 24.733 - type: recall_at_5 value: 32.385000000000005 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (sw) config: sw split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 21.401 - type: map_at_10 value: 34.205000000000005 - type: map_at_100 value: 35.571000000000005 - type: map_at_1000 value: 35.682 - type: map_at_20 value: 35.077000000000005 - type: map_at_3 value: 30.706 - type: map_at_5 value: 32.902 - type: mrr_at_1 value: 32.78 - type: mrr_at_10 value: 44.205 - type: mrr_at_100 value: 45.013 - type: mrr_at_1000 value: 45.055 - type: mrr_at_20 value: 44.726 - type: mrr_at_3 value: 41.909 - type: mrr_at_5 value: 43.434 - type: ndcg_at_1 value: 32.78 - type: ndcg_at_10 value: 41.385 - type: ndcg_at_100 value: 46.568 - type: ndcg_at_1000 value: 48.881 - type: ndcg_at_20 value: 43.872 - type: ndcg_at_3 value: 36.634 - type: ndcg_at_5 value: 38.964999999999996 - type: precision_at_1 value: 32.78 - type: precision_at_10 value: 8.859 - type: precision_at_100 value: 1.32 - type: precision_at_1000 value: 0.16199999999999998 - type: precision_at_20 value: 5.27 - type: precision_at_3 value: 20.608999999999998 - type: precision_at_5 value: 14.979000000000001 - type: recall_at_1 value: 21.401 - type: recall_at_10 value: 52.471000000000004 - type: recall_at_100 value: 72.069 - type: recall_at_1000 value: 87.996 - type: recall_at_20 value: 60.589999999999996 - type: recall_at_3 value: 39.28 - type: recall_at_5 value: 46.015 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (swahili) config: swahili split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 29.005 - type: map_at_10 value: 40.907 - type: map_at_100 value: 41.557 - type: map_at_1000 value: 41.604 - type: map_at_20 value: 41.234 - type: map_at_3 value: 38.114 - type: map_at_5 value: 39.6 - type: mrr_at_1 value: 30.597 - type: mrr_at_10 value: 42.0 - type: mrr_at_100 value: 42.557 - type: mrr_at_1000 value: 42.601 - type: mrr_at_20 value: 42.272999999999996 - type: mrr_at_3 value: 39.378 - type: mrr_at_5 value: 40.759 - type: ndcg_at_1 value: 30.597 - type: ndcg_at_10 value: 46.864 - type: ndcg_at_100 value: 50.099000000000004 - type: ndcg_at_1000 value: 51.354 - type: ndcg_at_20 value: 47.94 - type: ndcg_at_3 value: 41.234 - type: ndcg_at_5 value: 43.822 - type: precision_at_1 value: 30.597 - type: precision_at_10 value: 6.984999999999999 - type: precision_at_100 value: 0.8750000000000001 - type: precision_at_1000 value: 0.099 - type: precision_at_20 value: 3.7310000000000003 - type: precision_at_3 value: 17.463 - type: precision_at_5 value: 11.91 - type: recall_at_1 value: 29.005 - type: recall_at_10 value: 64.20400000000001 - type: recall_at_100 value: 79.403 - type: recall_at_1000 value: 89.104 - type: recall_at_20 value: 68.234 - type: recall_at_3 value: 49.055 - type: recall_at_5 value: 55.149 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (te) config: te split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 40.781 - type: map_at_10 value: 53.428 - type: map_at_100 value: 54.118 - type: map_at_1000 value: 54.135999999999996 - type: map_at_20 value: 53.873000000000005 - type: map_at_3 value: 50.205 - type: map_at_5 value: 52.356 - type: mrr_at_1 value: 41.184 - type: mrr_at_10 value: 53.74100000000001 - type: mrr_at_100 value: 54.388999999999996 - type: mrr_at_1000 value: 54.406 - type: mrr_at_20 value: 54.151 - type: mrr_at_3 value: 50.664 - type: mrr_at_5 value: 52.717000000000006 - type: ndcg_at_1 value: 41.184 - type: ndcg_at_10 value: 59.614999999999995 - type: ndcg_at_100 value: 62.875 - type: ndcg_at_1000 value: 63.368 - type: ndcg_at_20 value: 61.168 - type: ndcg_at_3 value: 53.322 - type: ndcg_at_5 value: 57.079 - type: precision_at_1 value: 41.184 - type: precision_at_10 value: 8.043 - type: precision_at_100 value: 0.963 - type: precision_at_1000 value: 0.1 - type: precision_at_20 value: 4.342 - type: precision_at_3 value: 20.934 - type: precision_at_5 value: 14.469000000000001 - type: recall_at_1 value: 40.781 - type: recall_at_10 value: 78.523 - type: recall_at_100 value: 93.639 - type: recall_at_1000 value: 97.504 - type: recall_at_20 value: 84.56099999999999 - type: recall_at_3 value: 61.895999999999994 - type: recall_at_5 value: 70.79299999999999 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (telugu) config: telugu split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 43.344 - type: map_at_10 value: 56.452000000000005 - type: map_at_100 value: 57.108000000000004 - type: map_at_1000 value: 57.131 - type: map_at_20 value: 56.86600000000001 - type: map_at_3 value: 54.154 - type: map_at_5 value: 55.57 - type: mrr_at_1 value: 44.427 - type: mrr_at_10 value: 57.16 - type: mrr_at_100 value: 57.764 - type: mrr_at_1000 value: 57.785 - type: mrr_at_20 value: 57.548 - type: mrr_at_3 value: 55.031 - type: mrr_at_5 value: 56.330999999999996 - type: ndcg_at_1 value: 44.427 - type: ndcg_at_10 value: 62.208 - type: ndcg_at_100 value: 65.33099999999999 - type: ndcg_at_1000 value: 65.96 - type: ndcg_at_20 value: 63.671 - type: ndcg_at_3 value: 57.68600000000001 - type: ndcg_at_5 value: 60.126999999999995 - type: precision_at_1 value: 44.427 - type: precision_at_10 value: 8.251 - type: precision_at_100 value: 0.98 - type: precision_at_1000 value: 0.10300000000000001 - type: precision_at_20 value: 4.42 - type: precision_at_3 value: 23.116999999999997 - type: precision_at_5 value: 15.076999999999998 - type: recall_at_1 value: 43.344 - type: recall_at_10 value: 79.10199999999999 - type: recall_at_100 value: 93.57600000000001 - type: recall_at_1000 value: 98.529 - type: recall_at_20 value: 84.83000000000001 - type: recall_at_3 value: 67.02799999999999 - type: recall_at_5 value: 72.833 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (th) config: th split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 27.747 - type: map_at_10 value: 43.586999999999996 - type: map_at_100 value: 45.256 - type: map_at_1000 value: 45.339 - type: map_at_20 value: 44.628 - type: map_at_3 value: 38.751999999999995 - type: map_at_5 value: 41.510000000000005 - type: mrr_at_1 value: 40.791 - type: mrr_at_10 value: 53.551 - type: mrr_at_100 value: 54.31 - type: mrr_at_1000 value: 54.32599999999999 - type: mrr_at_20 value: 54.057 - type: mrr_at_3 value: 50.637 - type: mrr_at_5 value: 52.525999999999996 - type: ndcg_at_1 value: 40.791 - type: ndcg_at_10 value: 52.144999999999996 - type: ndcg_at_100 value: 57.977000000000004 - type: ndcg_at_1000 value: 59.24 - type: ndcg_at_20 value: 54.864999999999995 - type: ndcg_at_3 value: 45.074 - type: ndcg_at_5 value: 48.504999999999995 - type: precision_at_1 value: 40.791 - type: precision_at_10 value: 11.350999999999999 - type: precision_at_100 value: 1.6039999999999999 - type: precision_at_1000 value: 0.178 - type: precision_at_20 value: 6.589 - type: precision_at_3 value: 25.239 - type: precision_at_5 value: 18.526999999999997 - type: recall_at_1 value: 27.747 - type: recall_at_10 value: 66.752 - type: recall_at_100 value: 89.51400000000001 - type: recall_at_1000 value: 97.485 - type: recall_at_20 value: 75.658 - type: recall_at_3 value: 48.393 - type: recall_at_5 value: 56.977 - task: type: retrieval dataset: type: mteb/mrtydi name: MTEB MrTydiRetrieval (thai) config: thai split: test revision: fc24a3ce8f09746410daee3d5cd823ff7a0675b7 metrics: - type: map_at_1 value: 28.810000000000002 - type: map_at_10 value: 41.803000000000004 - type: map_at_100 value: 42.818 - type: map_at_1000 value: 42.861 - type: map_at_20 value: 42.487 - type: map_at_3 value: 37.719 - type: map_at_5 value: 40.143 - type: mrr_at_1 value: 31.596999999999998 - type: mrr_at_10 value: 43.784 - type: mrr_at_100 value: 44.647 - type: mrr_at_1000 value: 44.681 - type: mrr_at_20 value: 44.385000000000005 - type: mrr_at_3 value: 40.266000000000005 - type: mrr_at_5 value: 42.266 - type: ndcg_at_1 value: 31.596999999999998 - type: ndcg_at_10 value: 48.874 - type: ndcg_at_100 value: 53.285000000000004 - type: ndcg_at_1000 value: 54.398 - type: ndcg_at_20 value: 51.188 - type: ndcg_at_3 value: 41.010000000000005 - type: ndcg_at_5 value: 45.054 - type: precision_at_1 value: 31.596999999999998 - type: precision_at_10 value: 7.747999999999999 - type: precision_at_100 value: 1.008 - type: precision_at_1000 value: 0.11100000000000002 - type: precision_at_20 value: 4.382 - type: precision_at_3 value: 17.619 - type: precision_at_5 value: 12.806999999999999 - type: recall_at_1 value: 28.810000000000002 - type: recall_at_10 value: 68.88 - type: recall_at_100 value: 88.263 - type: recall_at_1000 value: 96.765 - type: recall_at_20 value: 77.647 - type: recall_at_3 value: 48.137 - type: recall_at_5 value: 57.577 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (yo) config: yo split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 16.387 - type: map_at_10 value: 23.988 - type: map_at_100 value: 24.65 - type: map_at_1000 value: 24.725 - type: map_at_20 value: 24.310000000000002 - type: map_at_3 value: 21.23 - type: map_at_5 value: 23.093 - type: mrr_at_1 value: 17.647 - type: mrr_at_10 value: 26.144000000000002 - type: mrr_at_100 value: 26.751 - type: mrr_at_1000 value: 26.804 - type: mrr_at_20 value: 26.43 - type: mrr_at_3 value: 23.669 - type: mrr_at_5 value: 25.392 - type: ndcg_at_1 value: 17.647 - type: ndcg_at_10 value: 28.583 - type: ndcg_at_100 value: 31.790000000000003 - type: ndcg_at_1000 value: 33.705 - type: ndcg_at_20 value: 29.669 - type: ndcg_at_3 value: 23.271 - type: ndcg_at_5 value: 26.509 - type: precision_at_1 value: 17.647 - type: precision_at_10 value: 4.79 - type: precision_at_100 value: 0.655 - type: precision_at_1000 value: 0.08499999999999999 - type: precision_at_20 value: 2.6470000000000002 - type: precision_at_3 value: 10.363999999999999 - type: precision_at_5 value: 7.899000000000001 - type: recall_at_1 value: 16.387 - type: recall_at_10 value: 40.476 - type: recall_at_100 value: 55.11200000000001 - type: recall_at_1000 value: 69.538 - type: recall_at_20 value: 44.538 - type: recall_at_3 value: 26.540999999999997 - type: recall_at_5 value: 34.384 - task: type: retrieval dataset: type: mteb/miracl-hard-negatives name: MTEB MIRACLRetrievalHardNegatives (zh) config: zh split: dev revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb metrics: - type: map_at_1 value: 13.971 - type: map_at_10 value: 27.076 - type: map_at_100 value: 30.128 - type: map_at_1000 value: 30.25 - type: map_at_20 value: 28.731 - type: map_at_3 value: 21.029999999999998 - type: map_at_5 value: 23.769000000000002 - type: mrr_at_1 value: 26.717999999999996 - type: mrr_at_10 value: 40.331 - type: mrr_at_100 value: 41.448 - type: mrr_at_1000 value: 41.461999999999996 - type: mrr_at_20 value: 41.103 - type: mrr_at_3 value: 36.302 - type: mrr_at_5 value: 38.414 - type: ndcg_at_1 value: 26.717999999999996 - type: ndcg_at_10 value: 36.744 - type: ndcg_at_100 value: 47.361 - type: ndcg_at_1000 value: 48.869 - type: ndcg_at_20 value: 41.097 - type: ndcg_at_3 value: 28.322000000000003 - type: ndcg_at_5 value: 30.875999999999998 - type: precision_at_1 value: 26.717999999999996 - type: precision_at_10 value: 11.679 - type: precision_at_100 value: 2.2159999999999997 - type: precision_at_1000 value: 0.248 - type: precision_at_20 value: 7.545 - type: precision_at_3 value: 20.102 - type: precision_at_5 value: 16.131999999999998 - type: recall_at_1 value: 13.971 - type: recall_at_10 value: 50.763999999999996 - type: recall_at_100 value: 89.666 - type: recall_at_1000 value: 98.038 - type: recall_at_20 value: 64.1 - type: recall_at_3 value: 27.16 - type: recall_at_5 value: 35.022999999999996 widget: - source_sentence: ما هي أفضل الفنادق في ايبوهبالقرب من Ipoh Parade Shopping Centre؟ sentences: - Bei ORION gibt es eine Sale-Rubrik, in der alle reduzierten Artikel zu finden sind. Wenn du also auf der Suche nach einem Schnäppchen bist, weißt du, an welcher Stelle auf der Webseite du fündig wirst. Der Sale umfasst viele verschiedene Produke. Von Toys bis hin zu Dessous und Drogerieartikel - es spielt keine Rolle, wonach du suchst. Aufgrund der Produktvielfalt ist die Chance, dass du im Sale den passenden Gegenstand findest, groß. - عادة ما يكون لأصحاب النفوذ الجزئي ما بين 10000 و 100000 متابع. - المسافرون الموثّقون إلى مدينة ايبوه الذين أقاموا قرب Ipoh Parade Shopping Centre أعطوا أعلى التقييمات لـفندق فايل ، Zone Hotel (Ipoh) وGolden Roof Hotel Ampang Ipoh. - source_sentence: Habe ich Vorteile, wenn ich früh in das Projekt einsteige? sentences: - نعم لدينا خصومات مُتعددة على جميع أعمال السواتر والمظلات والجلسات ففي فصل الشتاء والصيف هناك خصومات مُتعددة في الأعياد والمناسبات - Der Vorteil einer frühen Mitgliedschaft besteht in der Möglichkeit der Mitgestaltung des Projektes. Alle später Hinzukommenden müssen die bis dahin getroffenen Entscheidungen akzeptieren. Zudem entscheidet unter anderem auch das Eintrittsdatum in die eG und das Engagement während des Projektverlaufes über die spätere Reihenfolge der Vergabe der Wohnungen. - Средняя оценка Registered от клиентов – 4 на основе 227 оценок и отзывов. Заходите на сайт и прочитайте реальные отзывы о Registered. - source_sentence: В какое время доступны ваши технические услуги? sentences: - Наша команда технической поддержки обеспечивает круглосуточное обслуживание в случае чрезвычайных ситуаций. Вы можете связаться с нами в любой день недели, в любое время суток и получить поддержку для ваших холодильных систем. Услуги по плановому техническому обслуживанию и ремонту предоставляются в обычное рабочее время, а услуги предоставляются в экстренных случаях, в том числе в ночное время и в выходные дни. - این سوال کاملا به علاقه و مهارت شما بستگی دارد. اگر به درس شیمی علاقه زیادی دارید این رشته بهترین انتخاب برای تحصیل در دانشگاه برای شما محسوب می‌شود. - 'eo光は10Gを提供している光回線です。 提供エリアは、通常プランと変わらず関西地方と福井県です。しかし、一部の利用できないエリアもあるので契約前に確認しましょう。 関連記事 eo光の10Gプランの評判口コミ' - source_sentence: Supertotobet redtiger oyun çeşitleri hangileri? sentences: - Supertotobet redtiger oyunları arasında gold star, golden tsar, golden lotus, blood suckers ve redtiger slot gibi çeşitli oyunlar vardır. Bu oyun seçeneklerini kullanabilmek için oyunlar hakkında bilgi sahibi olmalısınız. - Время выполнения проекта зависит от его сложности и размера. Обычно, время выполнения проекта составляет несколько месяцев. - 'Wer kein Homeoffice während des Coronavirus anbieten kann, ist dazu verpflichtet, Schutzmaßnahmen zu ergreifen. Die Arbeitsschutzbehörden der Länder sind befugt, Corona-Schutzmaßnahmen in Betrieben zu kontrollieren und Fehlverhalten zu bestrafen. Bei Verstößen sind Bußgelder in Höhe von bis zu 30.000 Euro möglich. Wiederholen sich schwere Verstöße, droht den Verantwortlichen sogar bis zu einem Jahr Freiheitsstrafe. Arbeitgeber, die die Vorschriften missachten, könnten zudem dafür haften, wenn Mitarbeiter durch eine Corona-Infektion gesundheitliche Schäden erleiden.' - source_sentence: Muss der Deckel der TipBox beim Autoklavieren geöffnet werden? sentences: - ВВП (валовый внутренний продукт) - это общая стоимость всех товаров и услуг, произведенных в стране за определенный период времени. Он является ключевым экономическим показателем, который отражает общий уровень экономической активности и роста. Инвесторы следят за ВВП, чтобы оценить состояние и перспективы экономики, потенциал для роста и возможности для инвестиций - برآمدگی های بیضه ممکن است نشان دهنده مشکلی در بیضه ها باشد. ممکن است به دلیل صدمه ای به وجود آمده یا ممکن است یک مشکل پزشکی جدی باشد. - Nein, das ist nicht notwendig. Die neue TipBox kann bei 121°C im geschlossenen Zustand autoklaviert werden. pipeline_tag: sentence-similarity library_name: sentence-transformers license: mit --- # SentenceTransformer This is a [sentence-transformers](https://www.SBERT.net) model trained. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [XLM-RoBERTa-base-MSMARCO](https://huggingface.co/PaDaS-Lab/xlm-roberta-base-msmarco) - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 dimensions - **Similarity Function:** Cosine Similarity - **Training Dataset:** [WebFAQ Retrieval Dataset](https://huggingface.co/datasets/PaDaS-Lab/webfaq-retrieval) - **License:** MIT ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: XLMRobertaModel (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, 'include_prompt': True}) ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("PaDaS-Lab/xlm-roberta-base-msmarco-webfaq") # Run inference sentences = [ 'Muss der Deckel der TipBox beim Autoklavieren geöffnet werden?', 'Nein, das ist nicht notwendig. Die neue TipBox kann bei 121°C im geschlossenen Zustand autoklaviert werden.', 'برآمدگی های بیضه ممکن است نشان دهنده مشکلی در بیضه ها باشد. ممکن است به دلیل صدمه ای به وجود آمده یا ممکن است یک مشکل پزشکی جدی باشد.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 2,560,000 training samples * Columns: sentence_0 and sentence_1 * Approximate statistics based on the first 1000 samples: | | sentence_0 | sentence_1 | |:--------|:----------------------------------------------------------------------------------|:------------------------------------------------------------------------------------| | type | string | string | | details | | | * Samples: | sentence_0 | sentence_1 | |:-----------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | Hat myTime ein großes Produktsortiment? | Das Sortiment von myTime umfasst mehr als 13.000 Lebensmittel. Du findest alle Produkte, die du auch im Supermarkt findest, darunter Obst und Gemüse, trockene Lebensmittel wie Pasta und Reis, Backwaren, Snacks und Tiefkühlkost. Auch Getränke wie Kaffee, Alkohol und Soda findest du im Online-Supermarkt. | | Gibt es eine Tigerspin App? | Tigerspin verzichtet auf eine mobile App. Wenn Sie ein paar Runden spielen möchten, öffnen Sie einfach die Webseite des Casinos und starten die Spiele im Browser. | | Bietet ihr auch maschinelle Übersetzungen an? Wenn ja, wann eignet sich diese und wann nicht? | Maschinelle Übersetzungen sind ein span­nen­des Thema, auch aktuell bei techtrans. Unter maschineller Übersetzung (MÜ) ver­steht man die auto­mati­sierte Über­tra­gung eines Aus­gangs­textes in die Ziel­sprache mittels einer so­ge­nannten Über­set­zungs­engine. Eine solche Engine kann nach re­gel­ba­sier­ten, statis­tischen oder neu­ro­nalen Prin­zipien auf­ge­baut sein.
Ob­wohl es maschinelle Über­set­zungs­engines schon seit einigen Jahr­zehn­ten gibt, ist erst mit der Ein­führung der neu­ro­nalen Engines (NMT) ca. ab dem Jahre 2015 die Output-Qualität ge­stie­gen. Nam­hafte Engine Provider sind zum Bei­spiel Google, DeepL, Microsoft, Amazon AWS und SDL. So ist es kaum ver­wunder­lich, dass diese Tech­no­logie zu­neh­mend Ein­zug sowohl in unseren All­tag als auch in den Über­set­zungs­pro­zess findet. | * Loss: [MultipleNegativesRankingLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters: ```json { "scale": 20.0, "similarity_fct": "cos_sim" } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `per_device_train_batch_size`: 128 - `per_device_eval_batch_size`: 128 - `num_train_epochs`: 1 - `fp16`: True - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters
Click to expand - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: no - `prediction_loss_only`: True - `per_device_train_batch_size`: 128 - `per_device_eval_batch_size`: 128 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `torch_empty_cache_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1 - `num_train_epochs`: 1 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.0 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: True - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: False - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: None - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `include_for_metrics`: [] - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin
### Training Logs | Epoch | Step | Training Loss | |:-----:|:-----:|:-------------:| | 0.025 | 500 | 0.1999 | | 0.05 | 1000 | 0.0279 | | 0.075 | 1500 | 0.0234 | | 0.1 | 2000 | 0.0203 | | 0.125 | 2500 | 0.0179 | | 0.15 | 3000 | 0.0171 | | 0.175 | 3500 | 0.0153 | | 0.2 | 4000 | 0.015 | | 0.225 | 4500 | 0.0143 | | 0.25 | 5000 | 0.014 | | 0.275 | 5500 | 0.0128 | | 0.3 | 6000 | 0.013 | | 0.325 | 6500 | 0.0129 | | 0.35 | 7000 | 0.0124 | | 0.375 | 7500 | 0.012 | | 0.4 | 8000 | 0.0121 | | 0.425 | 8500 | 0.0115 | | 0.45 | 9000 | 0.0113 | | 0.475 | 9500 | 0.0106 | | 0.5 | 10000 | 0.0107 | | 0.525 | 10500 | 0.011 | | 0.55 | 11000 | 0.0108 | | 0.575 | 11500 | 0.0103 | | 0.6 | 12000 | 0.0097 | | 0.625 | 12500 | 0.01 | | 0.65 | 13000 | 0.0104 | | 0.675 | 13500 | 0.0096 | | 0.7 | 14000 | 0.0096 | | 0.725 | 14500 | 0.0097 | | 0.75 | 15000 | 0.0097 | | 0.775 | 15500 | 0.0089 | | 0.8 | 16000 | 0.0089 | | 0.825 | 16500 | 0.0091 | | 0.85 | 17000 | 0.0085 | | 0.875 | 17500 | 0.0084 | | 0.9 | 18000 | 0.0089 | | 0.925 | 18500 | 0.0087 | | 0.95 | 19000 | 0.0087 | | 0.975 | 19500 | 0.0088 | | 1.0 | 20000 | 0.0089 | ### Framework Versions - Python: 3.11.5 - Sentence Transformers: 3.4.0 - Transformers: 4.48.0 - PyTorch: 2.5.1+cu124 - Accelerate: 1.2.1 - Datasets: 2.21.0 - Tokenizers: 0.21.0 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` #### WebFAQ Dataset ```bibtex @misc{dinzinger2025webfaq, title={WebFAQ: A Multilingual Collection of Natural Q&A Datasets for Dense Retrieval}, author={Michael Dinzinger and Laura Caspari and Kanishka Ghosh Dastidar and Jelena Mitrović and Michael Granitzer}, year={2025}, eprint={2502.20936}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```