--- tags: - sentence-transformers - feature-extraction - sentence-similarity - transformers - mteb license: apache-2.0 model-index: - name: bge-en-icl results: - dataset: config: en name: MTEB AmazonCounterfactualClassification (en) revision: e8379541af4e31359cca9fbcf4b00f2671dba205 split: test type: mteb/amazon_counterfactual metrics: - type: accuracy value: 93.1492537313433 - type: ap value: 72.56132559564212 - type: f1 value: 89.71796898040243 - type: main_score value: 93.1492537313433 task: type: Classification - dataset: config: default name: MTEB AmazonPolarityClassification revision: e2d317d38cd51312af73b3d32a06d1a08b442046 split: test type: mteb/amazon_polarity metrics: - type: accuracy value: 96.98372499999999 - type: ap value: 95.62303091773919 - type: f1 value: 96.98308191715637 - type: main_score value: 96.98372499999999 task: type: Classification - dataset: config: en name: MTEB AmazonReviewsClassification (en) revision: 1399c76144fd37290681b995c656ef9b2e06e26d split: test type: mteb/amazon_reviews_multi metrics: - type: accuracy value: 61.461999999999996 - type: f1 value: 60.57257766583118 - type: main_score value: 61.461999999999996 task: type: Classification - dataset: config: default name: MTEB ArguAna revision: c22ab2a51041ffd869aaddef7af8d8215647e41a split: test type: mteb/arguana metrics: - type: main_score value: 83.07967801208441 - type: ndcg_at_1 value: 66.50071123755335 - type: ndcg_at_3 value: 80.10869593172173 - type: ndcg_at_5 value: 81.89670542467924 - type: ndcg_at_10 value: 83.07967801208441 - type: ndcg_at_100 value: 83.5991349601075 - type: ndcg_at_1000 value: 83.5991349601075 - type: map_at_1 value: 66.50071123755335 - type: map_at_3 value: 76.83736367946898 - type: map_at_5 value: 77.8473210052158 - type: map_at_10 value: 78.35472690735851 - type: map_at_100 value: 78.47388207611678 - type: map_at_1000 value: 78.47388207611678 - type: precision_at_1 value: 66.50071123755335 - type: precision_at_3 value: 29.848269321953076 - type: precision_at_5 value: 18.762446657183045 - type: precision_at_10 value: 9.736842105262909 - type: precision_at_100 value: 0.9964438122332677 - type: precision_at_1000 value: 0.09964438122332549 - type: recall_at_1 value: 66.50071123755335 - type: recall_at_3 value: 89.5448079658606 - type: recall_at_5 value: 93.8122332859175 - type: recall_at_10 value: 97.36842105263158 - type: recall_at_100 value: 99.6443812233286 - type: recall_at_1000 value: 99.6443812233286 task: type: Retrieval - dataset: config: default name: MTEB ArxivClusteringP2P revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d split: test type: mteb/arxiv-clustering-p2p metrics: - type: main_score value: 54.43859683357485 - type: v_measure value: 54.43859683357485 - type: v_measure_std value: 14.511128158596337 task: type: Clustering - dataset: config: default name: MTEB ArxivClusteringS2S revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53 split: test type: mteb/arxiv-clustering-s2s metrics: - type: main_score value: 49.33365996236564 - type: v_measure value: 49.33365996236564 - type: v_measure_std value: 14.61261944856548 task: type: Clustering - dataset: config: default name: MTEB AskUbuntuDupQuestions revision: 2000358ca161889fa9c082cb41daa8dcfb161a54 split: test type: mteb/askubuntudupquestions-reranking metrics: - type: main_score value: 65.15263966490278 - type: map value: 65.15263966490278 - type: mrr value: 77.90331090885107 task: type: Reranking - dataset: config: default name: MTEB BIOSSES revision: d3fb88f8f02e40887cd149695127462bbcf29b4a split: test type: mteb/biosses-sts metrics: - type: main_score value: 86.47365710792691 - type: cosine_spearman value: 86.47365710792691 - type: spearman value: 86.47365710792691 task: type: STS - dataset: config: default name: MTEB Banking77Classification revision: 0fd18e25b25c072e09e0d92ab615fda904d66300 split: test type: mteb/banking77 metrics: - type: accuracy value: 91.48701298701299 - type: f1 value: 91.4733869423637 - type: main_score value: 91.48701298701299 task: type: Classification - dataset: config: default name: MTEB BiorxivClusteringP2P revision: 65b79d1d13f80053f67aca9498d9402c2d9f1f40 split: test type: mteb/biorxiv-clustering-p2p metrics: - type: main_score value: 53.050461108038036 - type: v_measure value: 53.050461108038036 - type: v_measure_std value: 0.9436104839012786 task: type: Clustering - dataset: config: default name: MTEB BiorxivClusteringS2S revision: 258694dd0231531bc1fd9de6ceb52a0853c6d908 split: test type: mteb/biorxiv-clustering-s2s metrics: - type: main_score value: 48.38215568371151 - type: v_measure value: 48.38215568371151 - type: v_measure_std value: 0.9104384504649026 task: type: Clustering - dataset: config: default name: MTEB CQADupstackRetrieval revision: 4ffe81d471b1924886b33c7567bfb200e9eec5c4 split: test type: mteb/cqadupstack metrics: - type: main_score value: 47.308084499970704 - type: ndcg_at_1 value: 36.038578730542476 - type: ndcg_at_3 value: 41.931365356453036 - type: ndcg_at_5 value: 44.479015523894994 - type: ndcg_at_10 value: 47.308084499970704 - type: ndcg_at_100 value: 52.498062430513606 - type: ndcg_at_1000 value: 54.2908789514719 - type: map_at_1 value: 30.38821701528966 - type: map_at_3 value: 37.974871761903636 - type: map_at_5 value: 39.85399878507757 - type: map_at_10 value: 41.31456611036795 - type: map_at_100 value: 42.62907836655835 - type: map_at_1000 value: 42.737235870659845 - type: precision_at_1 value: 36.038578730542476 - type: precision_at_3 value: 19.39960180094633 - type: precision_at_5 value: 13.79264655952497 - type: precision_at_10 value: 8.399223517333388 - type: precision_at_100 value: 1.2992373779520896 - type: precision_at_1000 value: 0.16327170951909567 - type: recall_at_1 value: 30.38821701528966 - type: recall_at_3 value: 45.51645512564165 - type: recall_at_5 value: 52.06077167834868 - type: recall_at_10 value: 60.38864106788279 - type: recall_at_100 value: 82.76968509918343 - type: recall_at_1000 value: 94.84170217080344 task: type: Retrieval - dataset: config: default name: MTEB ClimateFEVER revision: 47f2ac6acb640fc46020b02a5b59fdda04d39380 split: test type: mteb/climate-fever metrics: - type: main_score value: 45.4272998284769 - type: ndcg_at_1 value: 44.36482084690554 - type: ndcg_at_3 value: 38.13005747178844 - type: ndcg_at_5 value: 40.83474510717123 - type: ndcg_at_10 value: 45.4272998284769 - type: ndcg_at_100 value: 52.880220707479516 - type: ndcg_at_1000 value: 55.364753427333 - type: map_at_1 value: 19.200868621064064 - type: map_at_3 value: 28.33785740137525 - type: map_at_5 value: 31.67162504524064 - type: map_at_10 value: 34.417673164090075 - type: map_at_100 value: 36.744753097028976 - type: map_at_1000 value: 36.91262189016135 - type: precision_at_1 value: 44.36482084690554 - type: precision_at_3 value: 29.14223669923975 - type: precision_at_5 value: 22.410423452768388 - type: precision_at_10 value: 14.293159609120309 - type: precision_at_100 value: 2.248859934853431 - type: precision_at_1000 value: 0.2722475570032542 - type: recall_at_1 value: 19.200868621064064 - type: recall_at_3 value: 34.132464712269176 - type: recall_at_5 value: 42.35613463626491 - type: recall_at_10 value: 52.50814332247546 - type: recall_at_100 value: 77.16178067318128 - type: recall_at_1000 value: 90.59174809989138 task: type: Retrieval - dataset: config: default name: MTEB DBPedia revision: c0f706b76e590d620bd6618b3ca8efdd34e2d659 split: test type: mteb/dbpedia metrics: - type: main_score value: 51.634197691802754 - type: ndcg_at_1 value: 64.375 - type: ndcg_at_3 value: 55.677549598242614 - type: ndcg_at_5 value: 53.44347199908503 - type: ndcg_at_10 value: 51.634197691802754 - type: ndcg_at_100 value: 56.202861267183415 - type: ndcg_at_1000 value: 63.146019108272576 - type: map_at_1 value: 9.789380503780919 - type: map_at_3 value: 16.146582195277016 - type: map_at_5 value: 19.469695222167193 - type: map_at_10 value: 24.163327344766145 - type: map_at_100 value: 35.47047690245571 - type: map_at_1000 value: 37.5147432331838 - type: precision_at_1 value: 76.25 - type: precision_at_3 value: 59.08333333333333 - type: precision_at_5 value: 52.24999999999997 - type: precision_at_10 value: 42.54999999999994 - type: precision_at_100 value: 13.460000000000008 - type: precision_at_1000 value: 2.4804999999999966 - type: recall_at_1 value: 9.789380503780919 - type: recall_at_3 value: 17.48487134027656 - type: recall_at_5 value: 22.312024269698806 - type: recall_at_10 value: 30.305380335237324 - type: recall_at_100 value: 62.172868946596424 - type: recall_at_1000 value: 85.32410301328747 task: type: Retrieval - dataset: config: default name: MTEB EmotionClassification revision: 4f58c6b202a23cf9a4da393831edf4f9183cad37 split: test type: mteb/emotion metrics: - type: accuracy value: 93.36 - type: f1 value: 89.73665936982262 - type: main_score value: 93.36 task: type: Classification - dataset: config: default name: MTEB FEVER revision: bea83ef9e8fb933d90a2f1d5515737465d613e12 split: test type: mteb/fever metrics: - type: main_score value: 92.82809814626805 - type: ndcg_at_1 value: 88.98889888988899 - type: ndcg_at_3 value: 91.82404417747676 - type: ndcg_at_5 value: 92.41785792357787 - type: ndcg_at_10 value: 92.82809814626805 - type: ndcg_at_100 value: 93.31730867509245 - type: ndcg_at_1000 value: 93.45171203408582 - type: map_at_1 value: 82.64125817343636 - type: map_at_3 value: 89.39970782792554 - type: map_at_5 value: 89.96799501378695 - type: map_at_10 value: 90.27479706587437 - type: map_at_100 value: 90.45185655778057 - type: map_at_1000 value: 90.46130471574544 - type: precision_at_1 value: 88.98889888988899 - type: precision_at_3 value: 34.923492349234245 - type: precision_at_5 value: 21.524152415244043 - type: precision_at_10 value: 11.033603360337315 - type: precision_at_100 value: 1.1521152115211895 - type: precision_at_1000 value: 0.11765676567657675 - type: recall_at_1 value: 82.64125817343636 - type: recall_at_3 value: 94.35195900542428 - type: recall_at_5 value: 95.9071323799047 - type: recall_at_10 value: 97.04234113887586 - type: recall_at_100 value: 98.77282371094255 - type: recall_at_1000 value: 99.5555567461508 task: type: Retrieval - dataset: config: default name: MTEB FiQA2018 revision: 27a168819829fe9bcd655c2df245fb19452e8e06 split: test type: mteb/fiqa metrics: - type: main_score value: 59.67151242793314 - type: ndcg_at_1 value: 57.407407407407405 - type: ndcg_at_3 value: 53.79975378289304 - type: ndcg_at_5 value: 56.453379423655406 - type: ndcg_at_10 value: 59.67151242793314 - type: ndcg_at_100 value: 65.34055762539253 - type: ndcg_at_1000 value: 67.07707746043032 - type: map_at_1 value: 30.65887045053714 - type: map_at_3 value: 44.09107110881799 - type: map_at_5 value: 48.18573748068346 - type: map_at_10 value: 51.03680979612876 - type: map_at_100 value: 53.03165194566928 - type: map_at_1000 value: 53.16191096190861 - type: precision_at_1 value: 57.407407407407405 - type: precision_at_3 value: 35.493827160493886 - type: precision_at_5 value: 26.913580246913547 - type: precision_at_10 value: 16.435185185185155 - type: precision_at_100 value: 2.2685185185184986 - type: precision_at_1000 value: 0.25864197530863964 - type: recall_at_1 value: 30.65887045053714 - type: recall_at_3 value: 48.936723427464194 - type: recall_at_5 value: 58.55942925387371 - type: recall_at_10 value: 68.45128551147073 - type: recall_at_100 value: 88.24599311867836 - type: recall_at_1000 value: 98.18121693121691 task: type: Retrieval - dataset: config: default name: MTEB HotpotQA revision: ab518f4d6fcca38d87c25209f94beba119d02014 split: test type: mteb/hotpotqa metrics: - type: main_score value: 85.13780800141961 - type: ndcg_at_1 value: 89.9392302498312 - type: ndcg_at_3 value: 81.2061569376288 - type: ndcg_at_5 value: 83.53311592078133 - type: ndcg_at_10 value: 85.13780800141961 - type: ndcg_at_100 value: 87.02630661625386 - type: ndcg_at_1000 value: 87.47294723601075 - type: map_at_1 value: 44.9696151249156 - type: map_at_3 value: 76.46972766148966 - type: map_at_5 value: 78.47749268512187 - type: map_at_10 value: 79.49792611170005 - type: map_at_100 value: 80.09409086274644 - type: map_at_1000 value: 80.11950878917663 - type: precision_at_1 value: 89.9392302498312 - type: precision_at_3 value: 53.261309925724234 - type: precision_at_5 value: 33.79338284942924 - type: precision_at_10 value: 17.69750168805041 - type: precision_at_100 value: 1.9141120864280805 - type: precision_at_1000 value: 0.19721809588118133 - type: recall_at_1 value: 44.9696151249156 - type: recall_at_3 value: 79.8919648885888 - type: recall_at_5 value: 84.48345712356516 - type: recall_at_10 value: 88.48750844024308 - type: recall_at_100 value: 95.70560432140446 - type: recall_at_1000 value: 98.60904794058068 task: type: Retrieval - dataset: config: default name: MTEB ImdbClassification revision: 3d86128a09e091d6018b6d26cad27f2739fc2db7 split: test type: mteb/imdb metrics: - type: accuracy value: 96.9144 - type: ap value: 95.45276911068486 - type: f1 value: 96.91412729455966 - type: main_score value: 96.9144 task: type: Classification - dataset: config: default name: MTEB MSMARCO revision: c5a29a104738b98a9e76336939199e264163d4a0 split: dev type: mteb/msmarco metrics: - type: main_score value: 46.78865753107054 - type: ndcg_at_1 value: 26.63323782234957 - type: ndcg_at_3 value: 38.497585804985754 - type: ndcg_at_5 value: 42.72761631631636 - type: ndcg_at_10 value: 46.78865753107054 - type: ndcg_at_100 value: 51.96170786623209 - type: ndcg_at_1000 value: 52.82713901970963 - type: map_at_1 value: 25.89063992359121 - type: map_at_3 value: 35.299466730340654 - type: map_at_5 value: 37.68771887933786 - type: map_at_10 value: 39.40908074468253 - type: map_at_100 value: 40.53444082323405 - type: map_at_1000 value: 40.57183037649452 - type: precision_at_1 value: 26.63323782234957 - type: precision_at_3 value: 16.265520534861793 - type: precision_at_5 value: 11.902578796562304 - type: precision_at_10 value: 7.262177650430416 - type: precision_at_100 value: 0.9819484240687512 - type: precision_at_1000 value: 0.10571633237823287 - type: recall_at_1 value: 25.89063992359121 - type: recall_at_3 value: 46.99737344794652 - type: recall_at_5 value: 57.160936007640906 - type: recall_at_10 value: 69.43409742120343 - type: recall_at_100 value: 92.86413562559697 - type: recall_at_1000 value: 99.3230659025788 task: type: Retrieval - dataset: config: en name: MTEB MTOPDomainClassification (en) revision: d80d48c1eb48d3562165c59d59d0034df9fff0bf split: test type: mteb/mtop_domain metrics: - type: accuracy value: 98.42225262197901 - type: f1 value: 98.31652547061115 - type: main_score value: 98.42225262197901 task: type: Classification - dataset: config: en name: MTEB MTOPIntentClassification (en) revision: ae001d0e6b1228650b7bd1c2c65fb50ad11a8aba split: test type: mteb/mtop_intent metrics: - type: accuracy value: 94.00136798905609 - type: f1 value: 82.7022316533099 - type: main_score value: 94.00136798905609 task: type: Classification - dataset: config: en name: MTEB MassiveIntentClassification (en) revision: 4672e20407010da34463acc759c162ca9734bca6 split: test type: mteb/amazon_massive_intent metrics: - type: accuracy value: 82.92535305985204 - type: f1 value: 79.885538231847 - type: main_score value: 82.92535305985204 task: type: Classification - dataset: config: en name: MTEB MassiveScenarioClassification (en) revision: fad2c6e8459f9e1c45d9315f4953d921437d70f8 split: test type: mteb/amazon_massive_scenario metrics: - type: accuracy value: 85.60188298587758 - type: f1 value: 84.87416963499224 - type: main_score value: 85.60188298587758 task: type: Classification - dataset: config: default name: MTEB MedrxivClusteringP2P revision: e7a26af6f3ae46b30dde8737f02c07b1505bcc73 split: test type: mteb/medrxiv-clustering-p2p metrics: - type: main_score value: 45.86171497327639 - type: v_measure value: 45.86171497327639 - type: v_measure_std value: 1.551347259003324 task: type: Clustering - dataset: config: default name: MTEB MedrxivClusteringS2S revision: 35191c8c0dca72d8ff3efcd72aa802307d469663 split: test type: mteb/medrxiv-clustering-s2s metrics: - type: main_score value: 44.33336692345644 - type: v_measure value: 44.33336692345644 - type: v_measure_std value: 1.5931408596404715 task: type: Clustering - dataset: config: default name: MTEB MindSmallReranking revision: 59042f120c80e8afa9cdbb224f67076cec0fc9a7 split: test type: mteb/mind_small metrics: - type: main_score value: 30.597409734750503 - type: map value: 30.597409734750503 - type: mrr value: 31.397041548018457 task: type: Reranking - dataset: config: default name: MTEB NFCorpus revision: ec0fa4fe99da2ff19ca1214b7966684033a58814 split: test type: mteb/nfcorpus metrics: - type: main_score value: 41.850870119787835 - type: ndcg_at_1 value: 52.47678018575851 - type: ndcg_at_3 value: 47.43993801247414 - type: ndcg_at_5 value: 45.08173173082719 - type: ndcg_at_10 value: 41.850870119787835 - type: ndcg_at_100 value: 37.79284946590978 - type: ndcg_at_1000 value: 46.58046062123418 - type: map_at_1 value: 6.892464464226138 - type: map_at_3 value: 12.113195798233127 - type: map_at_5 value: 13.968475602788812 - type: map_at_10 value: 16.47564069781326 - type: map_at_100 value: 20.671726065190025 - type: map_at_1000 value: 22.328875914012006 - type: precision_at_1 value: 53.86996904024768 - type: precision_at_3 value: 43.96284829721363 - type: precision_at_5 value: 38.69969040247682 - type: precision_at_10 value: 30.928792569659457 - type: precision_at_100 value: 9.507739938080498 - type: precision_at_1000 value: 2.25882352941176 - type: recall_at_1 value: 6.892464464226138 - type: recall_at_3 value: 13.708153358278407 - type: recall_at_5 value: 16.651919797359145 - type: recall_at_10 value: 21.01801714352559 - type: recall_at_100 value: 37.01672102843443 - type: recall_at_1000 value: 69.8307270724072 task: type: Retrieval - dataset: config: default name: MTEB NQ revision: b774495ed302d8c44a3a7ea25c90dbce03968f31 split: test type: mteb/nq metrics: - type: main_score value: 73.88350836507092 - type: ndcg_at_1 value: 57.0683661645423 - type: ndcg_at_3 value: 67.89935813080585 - type: ndcg_at_5 value: 71.47769719452941 - type: ndcg_at_10 value: 73.88350836507092 - type: ndcg_at_100 value: 75.76561068060907 - type: ndcg_at_1000 value: 75.92437662684215 - type: map_at_1 value: 51.00424874468904 - type: map_at_3 value: 63.87359984550011 - type: map_at_5 value: 66.23696407879494 - type: map_at_10 value: 67.42415446608673 - type: map_at_100 value: 67.92692839842621 - type: map_at_1000 value: 67.93437922640133 - type: precision_at_1 value: 57.0683661645423 - type: precision_at_3 value: 29.692931633836416 - type: precision_at_5 value: 20.046349942062854 - type: precision_at_10 value: 10.950173812283 - type: precision_at_100 value: 1.1995944380069687 - type: precision_at_1000 value: 0.12146581691772171 - type: recall_at_1 value: 51.00424874468904 - type: recall_at_3 value: 75.93665507918116 - type: recall_at_5 value: 83.95133256083433 - type: recall_at_10 value: 90.78794901506375 - type: recall_at_100 value: 98.61915797605253 - type: recall_at_1000 value: 99.7827346465817 task: type: Retrieval - dataset: config: default name: MTEB QuoraRetrieval revision: e4e08e0b7dbe3c8700f0daef558ff32256715259 split: test type: mteb/quora metrics: - type: main_score value: 90.95410848372035 - type: ndcg_at_1 value: 84.61999999999999 - type: ndcg_at_3 value: 88.57366734033212 - type: ndcg_at_5 value: 89.89804048972175 - type: ndcg_at_10 value: 90.95410848372035 - type: ndcg_at_100 value: 91.83227134455773 - type: ndcg_at_1000 value: 91.88368412611601 - type: map_at_1 value: 73.4670089207039 - type: map_at_3 value: 84.87862925508942 - type: map_at_5 value: 86.68002324701408 - type: map_at_10 value: 87.7165466015312 - type: map_at_100 value: 88.28718809614146 - type: map_at_1000 value: 88.29877148480672 - type: precision_at_1 value: 84.61999999999999 - type: precision_at_3 value: 38.82333333333838 - type: precision_at_5 value: 25.423999999998642 - type: precision_at_10 value: 13.787999999998583 - type: precision_at_100 value: 1.5442999999999767 - type: precision_at_1000 value: 0.15672999999997972 - type: recall_at_1 value: 73.4670089207039 - type: recall_at_3 value: 89.98389854832143 - type: recall_at_5 value: 93.88541046010576 - type: recall_at_10 value: 96.99779417520634 - type: recall_at_100 value: 99.80318763957743 - type: recall_at_1000 value: 99.99638888888889 task: type: Retrieval - dataset: config: default name: MTEB RedditClustering revision: 24640382cdbf8abc73003fb0fa6d111a705499eb split: test type: mteb/reddit-clustering metrics: - type: main_score value: 72.33008348681277 - type: v_measure value: 72.33008348681277 - type: v_measure_std value: 2.9203215463933008 task: type: Clustering - dataset: config: default name: MTEB RedditClusteringP2P revision: 385e3cb46b4cfa89021f56c4380204149d0efe33 split: test type: mteb/reddit-clustering-p2p metrics: - type: main_score value: 72.72079657828903 - type: v_measure value: 72.72079657828903 - type: v_measure_std value: 11.930271663428735 task: type: Clustering - dataset: config: default name: MTEB SCIDOCS revision: f8c2fcf00f625baaa80f62ec5bd9e1fff3b8ae88 split: test type: mteb/scidocs metrics: - type: main_score value: 25.25865384510787 - type: ndcg_at_1 value: 28.7 - type: ndcg_at_3 value: 23.61736427940938 - type: ndcg_at_5 value: 20.845690325673885 - type: ndcg_at_10 value: 25.25865384510787 - type: ndcg_at_100 value: 36.18596641088721 - type: ndcg_at_1000 value: 41.7166868935345 - type: map_at_1 value: 5.828333333333361 - type: map_at_3 value: 10.689166666666676 - type: map_at_5 value: 13.069916666666668 - type: map_at_10 value: 15.4901164021164 - type: map_at_100 value: 18.61493245565425 - type: map_at_1000 value: 18.99943478016456 - type: precision_at_1 value: 28.7 - type: precision_at_3 value: 22.30000000000006 - type: precision_at_5 value: 18.55999999999997 - type: precision_at_10 value: 13.289999999999946 - type: precision_at_100 value: 2.905000000000005 - type: precision_at_1000 value: 0.4218999999999946 - type: recall_at_1 value: 5.828333333333361 - type: recall_at_3 value: 13.548333333333387 - type: recall_at_5 value: 18.778333333333308 - type: recall_at_10 value: 26.939999999999902 - type: recall_at_100 value: 58.91333333333344 - type: recall_at_1000 value: 85.57499999999972 task: type: Retrieval - dataset: config: default name: MTEB SICK-R revision: 20a6d6f312dd54037fe07a32d58e5e168867909d split: test type: mteb/sickr-sts metrics: - type: main_score value: 83.86733787791422 - type: cosine_spearman value: 83.86733787791422 - type: spearman value: 83.86733787791422 task: type: STS - dataset: config: default name: MTEB STS12 revision: a0d554a64d88156834ff5ae9920b964011b16384 split: test type: mteb/sts12-sts metrics: - type: main_score value: 78.14269330480724 - type: cosine_spearman value: 78.14269330480724 - type: spearman value: 78.14269330480724 task: type: STS - dataset: config: default name: MTEB STS13 revision: 7e90230a92c190f1bf69ae9002b8cea547a64cca split: test type: mteb/sts13-sts metrics: - type: main_score value: 86.58640009300751 - type: cosine_spearman value: 86.58640009300751 - type: spearman value: 86.58640009300751 task: type: STS - dataset: config: default name: MTEB STS14 revision: 6031580fec1f6af667f0bd2da0a551cf4f0b2375 split: test type: mteb/sts14-sts metrics: - type: main_score value: 82.8292579957437 - type: cosine_spearman value: 82.8292579957437 - type: spearman value: 82.8292579957437 task: type: STS - dataset: config: default name: MTEB STS15 revision: ae752c7c21bf194d8b67fd573edf7ae58183cbe3 split: test type: mteb/sts15-sts metrics: - type: main_score value: 87.77203714228862 - type: cosine_spearman value: 87.77203714228862 - type: spearman value: 87.77203714228862 task: type: STS - dataset: config: default name: MTEB STS16 revision: 4d8694f8f0e0100860b497b999b3dbed754a0513 split: test type: mteb/sts16-sts metrics: - type: main_score value: 87.0439304006969 - type: cosine_spearman value: 87.0439304006969 - type: spearman value: 87.0439304006969 task: type: STS - dataset: config: en-en name: MTEB STS17 (en-en) revision: faeb762787bd10488a50c8b5be4a3b82e411949c split: test type: mteb/sts17-crosslingual-sts metrics: - type: main_score value: 91.24736138013424 - type: cosine_spearman value: 91.24736138013424 - type: spearman value: 91.24736138013424 task: type: STS - dataset: config: en name: MTEB STS22 (en) revision: de9d86b3b84231dc21f76c7b7af1f28e2f57f6e3 split: test type: mteb/sts22-crosslingual-sts metrics: - type: main_score value: 70.07326214706 - type: cosine_spearman value: 70.07326214706 - type: spearman value: 70.07326214706 task: type: STS - dataset: config: default name: MTEB STSBenchmark revision: b0fddb56ed78048fa8b90373c8a3cfc37b684831 split: test type: mteb/stsbenchmark-sts metrics: - type: main_score value: 88.42076443255168 - type: cosine_spearman value: 88.42076443255168 - type: spearman value: 88.42076443255168 task: type: STS - dataset: config: default name: MTEB SciDocsRR revision: d3c5e1fc0b855ab6097bf1cda04dd73947d7caab split: test type: mteb/scidocs-reranking metrics: - type: main_score value: 86.9584489124583 - type: map value: 86.9584489124583 - type: mrr value: 96.59475328592976 task: type: Reranking - dataset: config: default name: MTEB SciFact revision: 0228b52cf27578f30900b9e5271d331663a030d7 split: test type: mteb/scifact metrics: - type: main_score value: 79.09159079425369 - type: ndcg_at_1 value: 66.0 - type: ndcg_at_3 value: 74.98853481223065 - type: ndcg_at_5 value: 77.29382051205019 - type: ndcg_at_10 value: 79.09159079425369 - type: ndcg_at_100 value: 80.29692802526776 - type: ndcg_at_1000 value: 80.55210036585547 - type: map_at_1 value: 62.994444444444454 - type: map_at_3 value: 71.7425925925926 - type: map_at_5 value: 73.6200925925926 - type: map_at_10 value: 74.50223544973547 - type: map_at_100 value: 74.82438594015447 - type: map_at_1000 value: 74.83420474892468 - type: precision_at_1 value: 66.0 - type: precision_at_3 value: 29.44444444444439 - type: precision_at_5 value: 19.40000000000008 - type: precision_at_10 value: 10.366666666666715 - type: precision_at_100 value: 1.0999999999999928 - type: precision_at_1000 value: 0.11200000000000007 - type: recall_at_1 value: 62.994444444444454 - type: recall_at_3 value: 80.89999999999998 - type: recall_at_5 value: 86.72777777777779 - type: recall_at_10 value: 91.88888888888887 - type: recall_at_100 value: 97.0 - type: recall_at_1000 value: 99.0 task: type: Retrieval - dataset: config: default name: MTEB SprintDuplicateQuestions revision: d66bd1f72af766a5cc4b0ca5e00c162f89e8cc46 split: test type: mteb/sprintduplicatequestions-pairclassification metrics: - type: main_score value: 97.26819027722253 - type: cos_sim_accuracy value: 99.88019801980198 - type: cos_sim_accuracy_threshold value: 76.67685151100159 - type: cos_sim_ap value: 97.23260568085786 - type: cos_sim_f1 value: 93.91824526420737 - type: cos_sim_f1_threshold value: 75.82710981369019 - type: cos_sim_precision value: 93.63817097415506 - type: cos_sim_recall value: 94.19999999999999 - type: dot_accuracy value: 99.88019801980198 - type: dot_accuracy_threshold value: 76.67686343193054 - type: dot_ap value: 97.23260568085786 - type: dot_f1 value: 93.91824526420737 - type: dot_f1_threshold value: 75.8271336555481 - type: dot_precision value: 93.63817097415506 - type: dot_recall value: 94.19999999999999 - type: euclidean_accuracy value: 99.88019801980198 - type: euclidean_accuracy_threshold value: 68.29807758331299 - type: euclidean_ap value: 97.23259982599497 - type: euclidean_f1 value: 93.91824526420737 - type: euclidean_f1_threshold value: 69.53110694885254 - type: euclidean_precision value: 93.63817097415506 - type: euclidean_recall value: 94.19999999999999 - type: manhattan_accuracy value: 99.87821782178217 - type: manhattan_accuracy_threshold value: 3482.6908111572266 - type: manhattan_ap value: 97.26819027722253 - type: manhattan_f1 value: 93.92592592592592 - type: manhattan_f1_threshold value: 3555.5641174316406 - type: manhattan_precision value: 92.78048780487805 - type: manhattan_recall value: 95.1 - type: max_accuracy value: 99.88019801980198 - type: max_ap value: 97.26819027722253 - type: max_f1 value: 93.92592592592592 task: type: PairClassification - dataset: config: default name: MTEB StackExchangeClustering revision: 6cbc1f7b2bc0622f2e39d2c77fa502909748c259 split: test type: mteb/stackexchange-clustering metrics: - type: main_score value: 81.32419328350603 - type: v_measure value: 81.32419328350603 - type: v_measure_std value: 2.666861121694755 task: type: Clustering - dataset: config: default name: MTEB StackExchangeClusteringP2P revision: 815ca46b2622cec33ccafc3735d572c266efdb44 split: test type: mteb/stackexchange-clustering-p2p metrics: - type: main_score value: 46.048387963107565 - type: v_measure value: 46.048387963107565 - type: v_measure_std value: 1.4102848576321703 task: type: Clustering - dataset: config: default name: MTEB StackOverflowDupQuestions revision: e185fbe320c72810689fc5848eb6114e1ef5ec69 split: test type: mteb/stackoverflowdupquestions-reranking metrics: - type: main_score value: 56.70574900554072 - type: map value: 56.70574900554072 - type: mrr value: 57.517109116373824 task: type: Reranking - dataset: config: default name: MTEB SummEval revision: cda12ad7615edc362dbf25a00fdd61d3b1eaf93c split: test type: mteb/summeval metrics: - type: main_score value: 30.76932903185174 - type: cosine_spearman value: 30.76932903185174 - type: spearman value: 30.76932903185174 task: type: Summarization - dataset: config: default name: MTEB TRECCOVID revision: bb9466bac8153a0349341eb1b22e06409e78ef4e split: test type: mteb/trec-covid metrics: - type: main_score value: 79.07987651251462 - type: ndcg_at_1 value: 83.0 - type: ndcg_at_3 value: 79.86598407528447 - type: ndcg_at_5 value: 79.27684428714952 - type: ndcg_at_10 value: 79.07987651251462 - type: ndcg_at_100 value: 64.55029164391163 - type: ndcg_at_1000 value: 59.42333857860492 - type: map_at_1 value: 0.226053732680979 - type: map_at_3 value: 0.644034626013194 - type: map_at_5 value: 1.045196967937728 - type: map_at_10 value: 2.0197496659905085 - type: map_at_100 value: 13.316018005224159 - type: map_at_1000 value: 33.784766957424104 - type: precision_at_1 value: 88.0 - type: precision_at_3 value: 86.66666666666667 - type: precision_at_5 value: 85.20000000000002 - type: precision_at_10 value: 84.19999999999997 - type: precision_at_100 value: 67.88000000000001 - type: precision_at_1000 value: 26.573999999999998 - type: recall_at_1 value: 0.226053732680979 - type: recall_at_3 value: 0.6754273711472734 - type: recall_at_5 value: 1.1168649828059245 - type: recall_at_10 value: 2.2215081031265207 - type: recall_at_100 value: 16.694165236664727 - type: recall_at_1000 value: 56.7022214857503 task: type: Retrieval - dataset: config: default name: MTEB Touche2020 revision: a34f9a33db75fa0cbb21bb5cfc3dae8dc8bec93f split: test type: mteb/touche2020 metrics: - type: main_score value: 30.47934263207554 - type: ndcg_at_1 value: 33.6734693877551 - type: ndcg_at_3 value: 34.36843900446739 - type: ndcg_at_5 value: 32.21323786731918 - type: ndcg_at_10 value: 30.47934263207554 - type: ndcg_at_100 value: 41.49598869753928 - type: ndcg_at_1000 value: 52.32963949183662 - type: map_at_1 value: 3.0159801678718168 - type: map_at_3 value: 7.13837927642557 - type: map_at_5 value: 9.274004610363466 - type: map_at_10 value: 12.957368366814324 - type: map_at_100 value: 19.3070585127604 - type: map_at_1000 value: 20.809777161133532 - type: precision_at_1 value: 34.69387755102041 - type: precision_at_3 value: 36.054421768707485 - type: precision_at_5 value: 32.24489795918368 - type: precision_at_10 value: 27.142857142857146 - type: precision_at_100 value: 8.326530612244898 - type: precision_at_1000 value: 1.5755102040816336 - type: recall_at_1 value: 3.0159801678718168 - type: recall_at_3 value: 8.321771388428257 - type: recall_at_5 value: 11.737532394366069 - type: recall_at_10 value: 19.49315139822179 - type: recall_at_100 value: 50.937064145519685 - type: recall_at_1000 value: 83.4358283484675 task: type: Retrieval - dataset: config: default name: MTEB ToxicConversationsClassification revision: edfaf9da55d3dd50d43143d90c1ac476895ae6de split: test type: mteb/toxic_conversations_50k metrics: - type: accuracy value: 93.173828125 - type: ap value: 46.040184641424396 - type: f1 value: 80.77280549412752 - type: main_score value: 93.173828125 task: type: Classification - dataset: config: default name: MTEB TweetSentimentExtractionClassification revision: d604517c81ca91fe16a244d1248fc021f9ecee7a split: test type: mteb/tweet_sentiment_extraction metrics: - type: accuracy value: 79.9320882852292 - type: f1 value: 80.22638685975485 - type: main_score value: 79.9320882852292 task: type: Classification - dataset: config: default name: MTEB TwentyNewsgroupsClustering revision: 6125ec4e24fa026cec8a478383ee943acfbd5449 split: test type: mteb/twentynewsgroups-clustering metrics: - type: main_score value: 68.98152919711418 - type: v_measure value: 68.98152919711418 - type: v_measure_std value: 1.2519720970652428 task: type: Clustering - dataset: config: default name: MTEB TwitterSemEval2015 revision: 70970daeab8776df92f5ea462b6173c0b46fd2d1 split: test type: mteb/twittersemeval2015-pairclassification metrics: - type: main_score value: 79.34189681158234 - type: cos_sim_accuracy value: 87.68552184538356 - type: cos_sim_accuracy_threshold value: 76.06316804885864 - type: cos_sim_ap value: 79.34189149773933 - type: cos_sim_f1 value: 72.16386554621849 - type: cos_sim_f1_threshold value: 73.62890243530273 - type: cos_sim_precision value: 71.82435964453737 - type: cos_sim_recall value: 72.5065963060686 - type: dot_accuracy value: 87.68552184538356 - type: dot_accuracy_threshold value: 76.06316208839417 - type: dot_ap value: 79.34189231911259 - type: dot_f1 value: 72.16386554621849 - type: dot_f1_threshold value: 73.62889647483826 - type: dot_precision value: 71.82435964453737 - type: dot_recall value: 72.5065963060686 - type: euclidean_accuracy value: 87.68552184538356 - type: euclidean_accuracy_threshold value: 69.19080018997192 - type: euclidean_ap value: 79.34189681158234 - type: euclidean_f1 value: 72.16386554621849 - type: euclidean_f1_threshold value: 72.62383103370667 - type: euclidean_precision value: 71.82435964453737 - type: euclidean_recall value: 72.5065963060686 - type: manhattan_accuracy value: 87.661679680515 - type: manhattan_accuracy_threshold value: 3408.807373046875 - type: manhattan_ap value: 79.29617544165136 - type: manhattan_f1 value: 72.1957671957672 - type: manhattan_f1_threshold value: 3597.7684020996094 - type: manhattan_precision value: 72.38726790450929 - type: manhattan_recall value: 72.00527704485488 - type: max_accuracy value: 87.68552184538356 - type: max_ap value: 79.34189681158234 - type: max_f1 value: 72.1957671957672 task: type: PairClassification - dataset: config: default name: MTEB TwitterURLCorpus revision: 8b6510b0b1fa4e4c4f879467980e9be563ec1cdf split: test type: mteb/twitterurlcorpus-pairclassification metrics: - type: main_score value: 87.8635519535718 - type: cos_sim_accuracy value: 89.80672953778088 - type: cos_sim_accuracy_threshold value: 73.09532165527344 - type: cos_sim_ap value: 87.84251379545145 - type: cos_sim_f1 value: 80.25858884373845 - type: cos_sim_f1_threshold value: 70.57080268859863 - type: cos_sim_precision value: 77.14103110353643 - type: cos_sim_recall value: 83.63874345549738 - type: dot_accuracy value: 89.80672953778088 - type: dot_accuracy_threshold value: 73.09532761573792 - type: dot_ap value: 87.84251881260793 - type: dot_f1 value: 80.25858884373845 - type: dot_f1_threshold value: 70.57079076766968 - type: dot_precision value: 77.14103110353643 - type: dot_recall value: 83.63874345549738 - type: euclidean_accuracy value: 89.80672953778088 - type: euclidean_accuracy_threshold value: 73.3548641204834 - type: euclidean_ap value: 87.84251335039049 - type: euclidean_f1 value: 80.25858884373845 - type: euclidean_f1_threshold value: 76.71923041343689 - type: euclidean_precision value: 77.14103110353643 - type: euclidean_recall value: 83.63874345549738 - type: manhattan_accuracy value: 89.78150347343501 - type: manhattan_accuracy_threshold value: 3702.7603149414062 - type: manhattan_ap value: 87.8635519535718 - type: manhattan_f1 value: 80.27105660516332 - type: manhattan_f1_threshold value: 3843.5962677001953 - type: manhattan_precision value: 76.9361101306036 - type: manhattan_recall value: 83.90822297505389 - type: max_accuracy value: 89.80672953778088 - type: max_ap value: 87.8635519535718 - type: max_f1 value: 80.27105660516332 task: type: PairClassification ---

FlagEmbedding

For more details please refer to our Github: [FlagEmbedding](https://github.com/FlagOpen/FlagEmbedding). **BGE-EN-ICL** primarily demonstrates the following capabilities: - In-context learning ability: By providing few-shot examples in the query, it can significantly enhance the model's ability to handle new tasks. - Outstanding performance: The model has achieved state-of-the-art (SOTA) performance on both BEIR and AIR-Bench. ## 📑 Open-source Plan - [x] Checkpoint - [ ] Training Data - [ ] Evaluation Pipeline - [ ] Technical Report We will release the technical report and training data for **BGE-EN-ICL** in the future. ## Usage ### Using FlagEmbedding ``` git clone https://github.com/FlagOpen/FlagEmbedding.git cd FlagEmbedding pip install -e . ``` ```python from FlagEmbedding import FlagICLModel queries = ["how much protein should a female eat", "summit define"] documents = [ "As a general guideline, the CDC's average requirement of protein for women ages 19 to 70 is 46 grams per day. But, as you can see from this chart, you'll need to increase that if you're expecting or training for a marathon. Check out the chart below to see how much protein you should be eating each day.", "Definition of summit for English Language Learners. : 1 the highest point of a mountain : the top of a mountain. : 2 the highest level. : 3 a meeting or series of meetings between the leaders of two or more governments." ] examples = [ {'instruct': 'Given a web search query, retrieve relevant passages that answer the query.', 'query': 'what is a virtual interface', 'response': "A virtual interface is a software-defined abstraction that mimics the behavior and characteristics of a physical network interface. It allows multiple logical network connections to share the same physical network interface, enabling efficient utilization of network resources. Virtual interfaces are commonly used in virtualization technologies such as virtual machines and containers to provide network connectivity without requiring dedicated hardware. They facilitate flexible network configurations and help in isolating network traffic for security and management purposes."}, {'instruct': 'Given a web search query, retrieve relevant passages that answer the query.', 'query': 'causes of back pain in female for a week', 'response': "Back pain in females lasting a week can stem from various factors. Common causes include muscle strain due to lifting heavy objects or improper posture, spinal issues like herniated discs or osteoporosis, menstrual cramps causing referred pain, urinary tract infections, or pelvic inflammatory disease. Pregnancy-related changes can also contribute. Stress and lack of physical activity may exacerbate symptoms. Proper diagnosis by a healthcare professional is crucial for effective treatment and management."} ] model = FlagICLModel('BAAI/bge-en-icl', query_instruction_for_retrieval="Given a web search query, retrieve relevant passages that answer the query.", examples_for_task=examples, # set `examples_for_task=None` to use model without examples use_fp16=True) # Setting use_fp16 to True speeds up computation with a slight performance degradation embeddings_1 = model.encode_queries(queries) embeddings_2 = model.encode_corpus(documents) similarity = embeddings_1 @ embeddings_2.T print(similarity) ``` By default, FlagICLModel will use all available GPUs when encoding. Please set `os.environ["CUDA_VISIBLE_DEVICES"]` to select specific GPUs. You also can set `os.environ["CUDA_VISIBLE_DEVICES"]=""` to make all GPUs unavailable. ### Using HuggingFace Transformers With the transformers package, you can use the model like this: First, you pass your input through the transformer model, then you select the last hidden state of the first token (i.e., [CLS]) as the sentence embedding. ```python import torch import torch.nn.functional as F from torch import Tensor from transformers import AutoTokenizer, AutoModel def last_token_pool(last_hidden_states: Tensor, attention_mask: Tensor) -> Tensor: left_padding = (attention_mask[:, -1].sum() == attention_mask.shape[0]) if left_padding: return last_hidden_states[:, -1] else: sequence_lengths = attention_mask.sum(dim=1) - 1 batch_size = last_hidden_states.shape[0] return last_hidden_states[torch.arange(batch_size, device=last_hidden_states.device), sequence_lengths] def get_detailed_instruct(task_description: str, query: str) -> str: return f'{task_description}\n{query}' def get_detailed_example(task_description: str, query: str, response: str) -> str: return f'{task_description}\n{query}\n{response}' def get_new_queries(queries, query_max_len, examples_prefix, tokenizer): inputs = tokenizer( queries, max_length=query_max_len - len(tokenizer('', add_special_tokens=False)['input_ids']) - len( tokenizer('\n', add_special_tokens=False)['input_ids']), return_token_type_ids=False, truncation=True, return_tensors=None, add_special_tokens=False ) prefix_ids = tokenizer(examples_prefix, add_special_tokens=False)['input_ids'] suffix_ids = tokenizer('\n', add_special_tokens=False)['input_ids'] new_max_length = (len(prefix_ids) + len(suffix_ids) + query_max_len + 8) // 8 * 8 + 8 new_queries = tokenizer.batch_decode(inputs['input_ids']) for i in range(len(new_queries)): new_queries[i] = examples_prefix + new_queries[i] + '\n' return new_max_length, new_queries task = 'Given a web search query, retrieve relevant passages that answer the query.' examples = [ {'instruct': 'Given a web search query, retrieve relevant passages that answer the query.', 'query': 'what is a virtual interface', 'response': "A virtual interface is a software-defined abstraction that mimics the behavior and characteristics of a physical network interface. It allows multiple logical network connections to share the same physical network interface, enabling efficient utilization of network resources. Virtual interfaces are commonly used in virtualization technologies such as virtual machines and containers to provide network connectivity without requiring dedicated hardware. They facilitate flexible network configurations and help in isolating network traffic for security and management purposes."}, {'instruct': 'Given a web search query, retrieve relevant passages that answer the query.', 'query': 'causes of back pain in female for a week', 'response': "Back pain in females lasting a week can stem from various factors. Common causes include muscle strain due to lifting heavy objects or improper posture, spinal issues like herniated discs or osteoporosis, menstrual cramps causing referred pain, urinary tract infections, or pelvic inflammatory disease. Pregnancy-related changes can also contribute. Stress and lack of physical activity may exacerbate symptoms. Proper diagnosis by a healthcare professional is crucial for effective treatment and management."} ] examples = [get_detailed_example(e['instruct'], e['query'], e['response']) for e in examples] examples_prefix = '\n\n'.join(examples) + '\n\n' # if there not exists any examples, just set examples_prefix = '' queries = [ get_detailed_instruct(task, 'how much protein should a female eat'), get_detailed_instruct(task, 'summit define') ] documents = [ "As a general guideline, the CDC's average requirement of protein for women ages 19 to 70 is 46 grams per day. But, as you can see from this chart, you'll need to increase that if you're expecting or training for a marathon. Check out the chart below to see how much protein you should be eating each day.", "Definition of summit for English Language Learners. : 1 the highest point of a mountain : the top of a mountain. : 2 the highest level. : 3 a meeting or series of meetings between the leaders of two or more governments." ] query_max_len, doc_max_len = 512, 512 tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-en-icl') model = AutoModel.from_pretrained('BAAI/bge-en-icl') model.eval() new_query_max_len, new_queries = get_new_queries(queries, query_max_len, examples_prefix, tokenizer) query_batch_dict = tokenizer(new_queries, max_length=new_query_max_len, padding=True, truncation=True, return_tensors='pt') doc_batch_dict = tokenizer(documents, max_length=doc_max_len, padding=True, truncation=True, return_tensors='pt') with torch.no_grad(): query_outputs = model(**query_batch_dict) query_embeddings = last_token_pool(query_outputs.last_hidden_state, query_batch_dict['attention_mask']) doc_outputs = model(**doc_batch_dict) doc_embeddings = last_token_pool(doc_outputs.last_hidden_state, doc_batch_dict['attention_mask']) # normalize embeddings query_embeddings = F.normalize(query_embeddings, p=2, dim=1) doc_embeddings = F.normalize(doc_embeddings, p=2, dim=1) scores = (query_embeddings @ doc_embeddings.T) * 100 print(scores.tolist()) ``` ## Evaluation `bge-en-icl` achieve **state-of-the-art performance on both MTEB and AIR-Bench leaderboard!** - **[MTEB](https://huggingface.co/spaces/mteb/leaderboard)**: ![BEIR](./results/MTEB.png) - **[BEIR](https://huggingface.co/spaces/mteb/leaderboard)**: ![BEIR](./results/BEIR.png) - **[AIR-Bench](https://huggingface.co/spaces/AIR-Bench/leaderboard)**: **QA (en, nDCG@10):** | AIR-Bench_24.04 | wiki | web | news | healthcare | law | finance | arxiv | msmarco | ALL (8) | | :--------------------------: | :-------: | :-------: | :-------: | :--------: | :-------: | :-------: | :-------: | :-------: | :-------: | | **e5-mistral-7b-instruct** | 61.67 | 44.41 | 48.18 | 56.32 | 19.32 | 54.79 | 44.78 | 59.03 | 48.56 | | **SFR-Embedding-Mistral** | 63.46 | 51.27 | 52.21 | 58.76 | 23.27 | 56.94 | 47.75 | 58.99 | 51.58 | | **NV-Embed-v1** | 62.84 | 50.42 | 51.46 | 58.53 | 20.65 | 49.89 | 46.10 | 60.27 | 50.02 | | **Linq-Embed-Mistral** | 61.04 | 48.41 | 49.44 | **60.18** | 20.34 | 50.04 | 47.56 | 60.50 | 49.69 | | **gte-Qwen2-7B-instruct** | 63.46 | 51.20 | 54.07 | 54.20 | 22.31 | **58.20** | 40.27 | 58.39 | 50.26 | | **stella_en_1.5B_v5** | 61.99 | 50.88 | 53.87 | 58.81 | 23.22 | 57.26 | 44.81 | 61.38 | 51.53 | | **bge-en-icl zero-shot** | 64.61 | 54.40 | 55.11 | 57.25 | 25.10 | 54.81 | 48.46 | 63.71 | 52.93 | | **bge-en-icl few-shot** | **64.94** | **55.11** | **56.02** | 58.85 | **28.29** | 57.16 | **50.04** | **64.50** | **54.36** | **Long-Doc (en, Recall@10):** | AIR-Bench_24.04 | arxiv (4) | book (2) | healthcare (5) | law (4) | ALL (15) | | :--------------------------: | :-------: | :-------: | :------------: | :-------: | :-------: | | **text-embedding-3-large** | 74.53 | 73.16 | 65.83 | 64.47 | 68.77 | | **e5-mistral-7b-instruct** | 72.14 | 72.44 | 68.44 | 62.92 | 68.49 | | **SFR-Embedding-Mistral** | 72.79 | 72.41 | 67.94 | 64.83 | 69.00 | | **NV-Embed-v1** | 77.65 | 75.49 | 72.38 | **69.55** | 73.45 | | **Linq-Embed-Mistral** | 75.46 | 73.81 | 71.58 | 68.58 | 72.11 | | **gte-Qwen2-7B-instruct** | 63.93 | 68.51 | 65.59 | 65.26 | 65.45 | | **stella_en_1.5B_v5** | 73.17 | 74.38 | 70.02 | 69.32 | 71.25 | | **bge-en-icl zero-shot** | 78.30 | 78.21 | 73.65 | 67.09 | 73.75 | | **bge-en-icl few-shot** | **79.63** | **79.36** | **74.80** | 67.79 | **74.83** | ## Model List `bge` is short for `BAAI general embedding`. | Model | Language | | Description | query instruction for retrieval [1] | |:--------------------------------------------------------------------------|:-------------------:|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:----------------------------------------------------------------------------------------------------------------------------------------------------:|:--------:| | [BAAI/bge-en-icl](https://huggingface.co/BAAI/bge-en-icl) | English | - | A LLM-based embedding model with in-context learning capabilities, which can fully leverage the model's potential based on a few shot examples | Provide instructions and few-shot examples freely based on the given task. | | [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) | Multilingual | [Inference](https://github.com/FlagOpen/FlagEmbedding/tree/master/FlagEmbedding/BGE_M3#usage) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/FlagEmbedding/BGE_M3) | Multi-Functionality(dense retrieval, sparse retrieval, multi-vector(colbert)), Multi-Linguality, and Multi-Granularity(8192 tokens) | | | [BAAI/llm-embedder](https://huggingface.co/BAAI/llm-embedder) | English | [Inference](./FlagEmbedding/llm_embedder/README.md) [Fine-tune](./FlagEmbedding/llm_embedder/README.md) | a unified embedding model to support diverse retrieval augmentation needs for LLMs | See [README](./FlagEmbedding/llm_embedder/README.md) | | [BAAI/bge-reranker-large](https://huggingface.co/BAAI/bge-reranker-large) | Chinese and English | [Inference](#usage-for-reranker) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/reranker) | a cross-encoder model which is more accurate but less efficient [2] | | | [BAAI/bge-reranker-base](https://huggingface.co/BAAI/bge-reranker-base) | Chinese and English | [Inference](#usage-for-reranker) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/reranker) | a cross-encoder model which is more accurate but less efficient [2] | | | [BAAI/bge-large-en-v1.5](https://huggingface.co/BAAI/bge-large-en-v1.5) | English | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | version 1.5 with more reasonable similarity distribution | `Represent this sentence for searching relevant passages: ` | | [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5) | English | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | version 1.5 with more reasonable similarity distribution | `Represent this sentence for searching relevant passages: ` | | [BAAI/bge-small-en-v1.5](https://huggingface.co/BAAI/bge-small-en-v1.5) | English | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | version 1.5 with more reasonable similarity distribution | `Represent this sentence for searching relevant passages: ` | | [BAAI/bge-large-zh-v1.5](https://huggingface.co/BAAI/bge-large-zh-v1.5) | Chinese | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | version 1.5 with more reasonable similarity distribution | `为这个句子生成表示以用于检索相关文章：` | | [BAAI/bge-base-zh-v1.5](https://huggingface.co/BAAI/bge-base-zh-v1.5) | Chinese | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | version 1.5 with more reasonable similarity distribution | `为这个句子生成表示以用于检索相关文章：` | | [BAAI/bge-small-zh-v1.5](https://huggingface.co/BAAI/bge-small-zh-v1.5) | Chinese | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | version 1.5 with more reasonable similarity distribution | `为这个句子生成表示以用于检索相关文章：` | | [BAAI/bge-large-en](https://huggingface.co/BAAI/bge-large-en) | English | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | :trophy: rank **1st** in [MTEB](https://huggingface.co/spaces/mteb/leaderboard) leaderboard | `Represent this sentence for searching relevant passages: ` | | [BAAI/bge-base-en](https://huggingface.co/BAAI/bge-base-en) | English | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | a base-scale model but with similar ability to `bge-large-en` | `Represent this sentence for searching relevant passages: ` | | [BAAI/bge-small-en](https://huggingface.co/BAAI/bge-small-en) | English | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | a small-scale model but with competitive performance | `Represent this sentence for searching relevant passages: ` | | [BAAI/bge-large-zh](https://huggingface.co/BAAI/bge-large-zh) | Chinese | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | :trophy: rank **1st** in [C-MTEB](https://github.com/FlagOpen/FlagEmbedding/tree/master/C_MTEB) benchmark | `为这个句子生成表示以用于检索相关文章：` | | [BAAI/bge-base-zh](https://huggingface.co/BAAI/bge-base-zh) | Chinese | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | a base-scale model but with similar ability to `bge-large-zh` | `为这个句子生成表示以用于检索相关文章：` | | [BAAI/bge-small-zh](https://huggingface.co/BAAI/bge-small-zh) | Chinese | [Inference](#usage-for-embedding-model) [Fine-tune](https://github.com/FlagOpen/FlagEmbedding/tree/master/examples/finetune) | a small-scale model but with competitive performance | `为这个句子生成表示以用于检索相关文章：` | ## Citation If you find this repository useful, please consider giving a star :star: and citation ``` @misc{bge_embedding, title={C-Pack: Packaged Resources To Advance General Chinese Embedding}, author={Shitao Xiao and Zheng Liu and Peitian Zhang and Niklas Muennighoff}, year={2023}, eprint={2309.07597}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## License FlagEmbedding is licensed under the [MIT License](https://github.com/FlagOpen/FlagEmbedding/blob/master/LICENSE).