Linq-Embed-Mistral-bnb-4bitオープンソース埋め込みモデル - テキスト分類、検索、クラスタリングタスクを効率的に完了する

ホーム

Linq Embed Mistral Bnb 4bit

ashercn97によって開発

Linq-Embed-MistralはMistralアーキテクチャに基づく埋め込みモデルで、テキスト分類、検索、クラスタリングタスクに特化しており、複数のMTEBベンチマークで優れた性能を発揮します。

テキスト埋め込み

Transformers

英語#テキスト埋め込み #マルチタスク学習 #高精度分類

ダウンロード数 147

リリース時間 : 4/10/2025

モデル概要

このモデルは主に高品質なテキスト埋め込みを生成するために使用され、テキスト分類、情報検索、文書クラスタリングなど、さまざまな自然言語処理タスクに適用できます。

モデル特徴

優れたマルチタスク性能

分類、検索、クラスタリングなど、さまざまな自然言語処理タスクで優れた性能を発揮

広範なベンチマークカバレッジ

MTEBの複数のベンチマークデータセットで包括的な評価を実施

効率的な検索能力

情報検索タスクで高い精度と再現率を達成

モデル能力

テキスト分類

情報検索

文書クラスタリング

意味的類似度計算

テキスト再ランキング

使用事例

電子商取引

製品レビューの感情分析

Amazon製品レビューの感情傾向を分析

Amazon極性分類タスクで95.70%の精度を達成

製品分類

Amazon製品を複数カテゴリに分類

Amazon多カテゴリレビュー分類で57.64%の精度を達成

金融

銀行カスタマーサポート分類

銀行の顧客問い合わせを分類

Banking77データセットで87.88%の精度を達成

情報検索

質問応答システム

質問応答システムで関連文書を検索

HotpotQAデータセットで70.08%の平均精度を達成

事実確認

ある主張を支持または反駁する証拠を検索

ClimateFEVERデータセットで31.50%の平均精度を達成

base_model:

Linq-AI-Research/Linq-Embed-Mistral tags:
bnb-my-repo
mteb
transformers
sentence-transformers model-index:
name: Linq-Embed-Mistral results:
- task: type: Classification dataset: type: mteb/amazon_counterfactual name: MTEB AmazonCounterfactualClassification (en) config: en split: test revision: e8379541af4e31359cca9fbcf4b00f2671dba205 metrics:
  - type: accuracy value: 84.43283582089552
  - type: ap value: 50.39222584035829
  - type: f1 value: 78.47906270064071
- task: type: Classification dataset: type: mteb/amazon_polarity name: MTEB AmazonPolarityClassification config: default split: test revision: e2d317d38cd51312af73b3d32a06d1a08b442046 metrics:
  - type: accuracy value: 95.70445
  - type: ap value: 94.28273900595173
  - type: f1 value: 95.70048412173735
- task: type: Classification dataset: type: mteb/amazon_reviews_multi name: MTEB AmazonReviewsClassification (en) config: en split: test revision: 1399c76144fd37290681b995c656ef9b2e06e26d metrics:
  - type: accuracy value: 57.644000000000005
  - type: f1 value: 56.993648296704876
- task: type: Retrieval dataset: type: mteb/arguana name: MTEB ArguAna config: default split: test revision: c22ab2a51041ffd869aaddef7af8d8215647e41a metrics:
  - type: map_at_1 value: 45.804
  - type: map_at_10 value: 61.742
  - type: map_at_100 value: 62.07899999999999
  - type: map_at_1000 value: 62.08
  - type: map_at_3 value: 57.717
  - type: map_at_5 value: 60.27
  - type: mrr_at_1 value: 47.226
  - type: mrr_at_10 value: 62.256
  - type: mrr_at_100 value: 62.601
  - type: mrr_at_1000 value: 62.601
  - type: mrr_at_3 value: 58.203
  - type: mrr_at_5 value: 60.767
  - type: ndcg_at_1 value: 45.804
  - type: ndcg_at_10 value: 69.649
  - type: ndcg_at_100 value: 70.902
  - type: ndcg_at_1000 value: 70.91199999999999
  - type: ndcg_at_3 value: 61.497
  - type: ndcg_at_5 value: 66.097
  - type: precision_at_1 value: 45.804
  - type: precision_at_10 value: 9.452
  - type: precision_at_100 value: 0.996
  - type: precision_at_1000 value: 0.1
  - type: precision_at_3 value: 24.135
  - type: precision_at_5 value: 16.714000000000002
  - type: recall_at_1 value: 45.804
  - type: recall_at_10 value: 94.523
  - type: recall_at_100 value: 99.57300000000001
  - type: recall_at_1000 value: 99.644
  - type: recall_at_3 value: 72.404
  - type: recall_at_5 value: 83.57
- task: type: Clustering dataset: type: mteb/arxiv-clustering-p2p name: MTEB ArxivClusteringP2P config: default split: test revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d metrics:
  - type: v_measure value: 51.47612678878609
- task: type: Clustering dataset: type: mteb/arxiv-clustering-s2s name: MTEB ArxivClusteringS2S config: default split: test revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53 metrics:
  - type: v_measure value: 47.2977392340418
- task: type: Reranking dataset: type: mteb/askubuntudupquestions-reranking name: MTEB AskUbuntuDupQuestions config: default split: test revision: 2000358ca161889fa9c082cb41daa8dcfb161a54 metrics:
  - type: map value: 66.82016765243456
  - type: mrr value: 79.55227982236292
- task: type: STS dataset: type: mteb/biosses-sts name: MTEB BIOSSES config: default split: test revision: d3fb88f8f02e40887cd149695127462bbcf29b4a metrics:
  - type: cos_sim_pearson value: 89.15068664186332
  - type: cos_sim_spearman value: 86.4013663041054
  - type: euclidean_pearson value: 87.36391302921588
  - type: euclidean_spearman value: 86.4013663041054
  - type: manhattan_pearson value: 87.46116676558589
  - type: manhattan_spearman value: 86.78149544753352
- task: type: Classification dataset: type: mteb/banking77 name: MTEB Banking77Classification config: default split: test revision: 0fd18e25b25c072e09e0d92ab615fda904d66300 metrics:
  - type: accuracy value: 87.88311688311688
  - type: f1 value: 87.82368154811464
- task: type: Clustering dataset: type: mteb/biorxiv-clustering-p2p name: MTEB BiorxivClusteringP2P config: default split: test revision: 65b79d1d13f80053f67aca9498d9402c2d9f1f40 metrics:
  - type: v_measure value: 42.72860396750569
- task: type: Clustering dataset: type: mteb/biorxiv-clustering-s2s name: MTEB BiorxivClusteringS2S config: default split: test revision: 258694dd0231531bc1fd9de6ceb52a0853c6d908 metrics:
  - type: v_measure value: 39.58412067938718
- task: type: Retrieval dataset: type: mteb/cqadupstack name: MTEB CQADupstackRetrieval config: default split: test revision: 4ffe81d471b1924886b33c7567bfb200e9eec5c4 metrics:
  - type: map_at_1 value: 30.082666666666665
  - type: map_at_10 value: 41.13875
  - type: map_at_100 value: 42.45525
  - type: map_at_1000 value: 42.561249999999994
  - type: map_at_3 value: 37.822750000000006
  - type: map_at_5 value: 39.62658333333333
  - type: mrr_at_1 value: 35.584
  - type: mrr_at_10 value: 45.4675
  - type: mrr_at_100 value: 46.31016666666667
  - type: mrr_at_1000 value: 46.35191666666666
  - type: mrr_at_3 value: 42.86674999999999
  - type: mrr_at_5 value: 44.31341666666666
  - type: ndcg_at_1 value: 35.584
  - type: ndcg_at_10 value: 47.26516666666667
  - type: ndcg_at_100 value: 52.49108333333332
  - type: ndcg_at_1000 value: 54.24575
  - type: ndcg_at_3 value: 41.83433333333334
  - type: ndcg_at_5 value: 44.29899999999999
  - type: precision_at_1 value: 35.584
  - type: precision_at_10 value: 8.390333333333334
  - type: precision_at_100 value: 1.2941666666666667
  - type: precision_at_1000 value: 0.16308333333333336
  - type: precision_at_3 value: 19.414583333333333
  - type: precision_at_5 value: 13.751
  - type: recall_at_1 value: 30.082666666666665
  - type: recall_at_10 value: 60.88875
  - type: recall_at_100 value: 83.35141666666667
  - type: recall_at_1000 value: 95.0805
  - type: recall_at_3 value: 45.683749999999996
  - type: recall_at_5 value: 52.08208333333333
- task: type: Retrieval dataset: type: mteb/climate-fever name: MTEB ClimateFEVER config: default split: test revision: 47f2ac6acb640fc46020b02a5b59fdda04d39380 metrics:
  - type: map_at_1 value: 16.747
  - type: map_at_10 value: 29.168
  - type: map_at_100 value: 31.304
  - type: map_at_1000 value: 31.496000000000002
  - type: map_at_3 value: 24.57
  - type: map_at_5 value: 26.886
  - type: mrr_at_1 value: 37.524
  - type: mrr_at_10 value: 50.588
  - type: mrr_at_100 value: 51.28
  - type: mrr_at_1000 value: 51.29899999999999
  - type: mrr_at_3 value: 47.438
  - type: mrr_at_5 value: 49.434
  - type: ndcg_at_1 value: 37.524
  - type: ndcg_at_10 value: 39.11
  - type: ndcg_at_100 value: 46.373999999999995
  - type: ndcg_at_1000 value: 49.370999999999995
  - type: ndcg_at_3 value: 32.964
  - type: ndcg_at_5 value: 35.028
  - type: precision_at_1 value: 37.524
  - type: precision_at_10 value: 12.137
  - type: precision_at_100 value: 1.9929999999999999
  - type: precision_at_1000 value: 0.256
  - type: precision_at_3 value: 24.886
  - type: precision_at_5 value: 18.762
  - type: recall_at_1 value: 16.747
  - type: recall_at_10 value: 45.486
  - type: recall_at_100 value: 69.705
  - type: recall_at_1000 value: 86.119
  - type: recall_at_3 value: 30.070999999999998
  - type: recall_at_5 value: 36.565
- task: type: Retrieval dataset: type: mteb/dbpedia name: MTEB DBPedia config: default split: test revision: c0f706b76e590d620bd6618b3ca8efdd34e2d659 metrics:
  - type: map_at_1 value: 10.495000000000001
  - type: map_at_10 value: 24.005000000000003
  - type: map_at_100 value: 34.37
  - type: map_at_1000 value: 36.268
  - type: map_at_3 value: 16.694
  - type: map_at_5 value: 19.845
  - type: mrr_at_1 value: 75.5
  - type: mrr_at_10 value: 82.458
  - type: mrr_at_100 value: 82.638
  - type: mrr_at_1000 value: 82.64
  - type: mrr_at_3 value: 81.25
  - type: mrr_at_5 value: 82.125
  - type: ndcg_at_1 value: 64.625
  - type: ndcg_at_10 value: 51.322
  - type: ndcg_at_100 value: 55.413999999999994
  - type: ndcg_at_1000 value: 62.169
  - type: ndcg_at_3 value: 56.818999999999996
  - type: ndcg_at_5 value: 54.32900000000001
  - type: precision_at_1 value: 75.5
  - type: precision_at_10 value: 40.849999999999994
  - type: precision_at_100 value: 12.882
  - type: precision_at_1000 value: 2.394
  - type: precision_at_3 value: 59.667
  - type: precision_at_5 value: 52.2
  - type: recall_at_1 value: 10.495000000000001
  - type: recall_at_10 value: 29.226000000000003
  - type: recall_at_100 value: 59.614
  - type: recall_at_1000 value: 81.862
  - type: recall_at_3 value: 17.97
  - type: recall_at_5 value: 22.438
- task: type: Classification dataset: type: mteb/emotion name: MTEB EmotionClassification config: default split: test revision: 4f58c6b202a23cf9a4da393831edf4f9183cad37 metrics:
  - type: accuracy value: 51.82
  - type: f1 value: 47.794956731921054
- task: type: Retrieval dataset: type: mteb/fever name: MTEB FEVER config: default split: test revision: bea83ef9e8fb933d90a2f1d5515737465d613e12 metrics:
  - type: map_at_1 value: 82.52199999999999
  - type: map_at_10 value: 89.794
  - type: map_at_100 value: 89.962
  - type: map_at_1000 value: 89.972
  - type: map_at_3 value: 88.95100000000001
  - type: map_at_5 value: 89.524
  - type: mrr_at_1 value: 88.809
  - type: mrr_at_10 value: 93.554
  - type: mrr_at_100 value: 93.577
  - type: mrr_at_1000 value: 93.577
  - type: mrr_at_3 value: 93.324
  - type: mrr_at_5 value: 93.516
  - type: ndcg_at_1 value: 88.809
  - type: ndcg_at_10 value: 92.419
  - type: ndcg_at_100 value: 92.95
  - type: ndcg_at_1000 value: 93.10000000000001
  - type: ndcg_at_3 value: 91.45299999999999
  - type: ndcg_at_5 value: 92.05
  - type: precision_at_1 value: 88.809
  - type: precision_at_10 value: 10.911999999999999
  - type: precision_at_100 value: 1.143
  - type: precision_at_1000 value: 0.117
  - type: precision_at_3 value: 34.623
  - type: precision_at_5 value: 21.343999999999998
  - type: recall_at_1 value: 82.52199999999999
  - type: recall_at_10 value: 96.59400000000001
  - type: recall_at_100 value: 98.55699999999999
  - type: recall_at_1000 value: 99.413
  - type: recall_at_3 value: 94.02199999999999
  - type: recall_at_5 value: 95.582
- task: type: Retrieval dataset: type: mteb/fiqa name: MTEB FiQA2018 config: default split: test revision: 27a168819829fe9bcd655c2df245fb19452e8e06 metrics:
  - type: map_at_1 value: 32.842
  - type: map_at_10 value: 53.147
  - type: map_at_100 value: 55.265
  - type: map_at_1000 value: 55.37
  - type: map_at_3 value: 46.495
  - type: map_at_5 value: 50.214999999999996
  - type: mrr_at_1 value: 61.574
  - type: mrr_at_10 value: 68.426
  - type: mrr_at_100 value: 68.935
  - type: mrr_at_1000 value: 68.95400000000001
  - type: mrr_at_3 value: 66.307
  - type: mrr_at_5 value: 67.611
  - type: ndcg_at_1 value: 61.574
  - type: ndcg_at_10 value: 61.205
  - type: ndcg_at_100 value: 67.25999999999999
  - type: ndcg_at_1000 value: 68.657
  - type: ndcg_at_3 value: 56.717
  - type: ndcg_at_5 value: 58.196999999999996
  - type: precision_at_1 value: 61.574
  - type: precision_at_10 value: 16.852
  - type: precision_at_100 value: 2.33
  - type: precision_at_1000 value: 0.256
  - type: precision_at_3 value: 37.5
  - type: precision_at_5 value: 27.468999999999998
  - type: recall_at_1 value: 32.842
  - type: recall_at_10 value: 68.157
  - type: recall_at_100 value: 89.5
  - type: recall_at_1000 value: 97.68599999999999
  - type: recall_at_3 value: 50.783
  - type: recall_at_5 value: 58.672000000000004
- task: type: Retrieval dataset: type: mteb/hotpotqa name: MTEB HotpotQA config: default split: test revision: ab518f4d6fcca38d87c25209f94beba119d02014 metrics:
  - type: map_at_1 value: 39.068000000000005
  - type: map_at_10 value: 69.253
  - type: map_at_100 value: 70.036
  - type: map_at_1000 value: 70.081
  - type: map_at_3 value: 65.621
  - type: map_at_5 value: 67.976
  - type: mrr_at_1 value: 78.13600000000001
  - type: mrr_at_10 value: 84.328
  - type: mrr_at_100 value: 84.515
  - type: mrr_at_1000 value: 84.52300000000001
  - type: mrr_at_3 value: 83.52199999999999
  - type: mrr_at_5 value: 84.019
  - type: ndcg_at_1 value: 78.13600000000001
  - type: ndcg_at_10 value: 76.236
  - type: ndcg_at_100 value: 78.891
  - type: ndcg_at_1000 value: 79.73400000000001
  - type: ndcg_at_3 value: 71.258
  - type: ndcg_at_5 value: 74.129
  - type: precision_at_1 value: 78.13600000000001
  - type: precision_at_10 value: 16.347
  - type: precision_at_100 value: 1.839
  - type: precision_at_1000 value: 0.19499999999999998
  - type: precision_at_3 value: 47.189
  - type: precision_at_5 value: 30.581999999999997
  - type: recall_at_1 value: 39.068000000000005
  - type: recall_at_10 value: 81.735
  - type: recall_at_100 value: 91.945
  - type: recall_at_1000 value: 97.44800000000001
  - type: recall_at_3 value: 70.783
  - type: recall_at_5 value: 76.455
- task: type: Classification dataset: type: mteb/imdb name: MTEB ImdbClassification config: default split: test revision: 3d86128a09e091d6018b6d26cad27f2739fc2db7 metrics:
  - type: accuracy value: 94.7764
  - type: ap value: 92.67841294818406
  - type: f1 value: 94.77375157383646
- task: type: Retrieval dataset: type: mteb/msmarco name: MTEB MSMARCO config: default split: dev revision: c5a29a104738b98a9e76336939199e264163d4a0 metrics:
  - type: map_at_1 value: 24.624
  - type: map_at_10 value: 37.861
  - type: map_at_100 value: 39.011
  - type: map_at_1000 value: 39.052
  - type: map_at_3 value: 33.76
  - type: map_at_5 value: 36.153
  - type: mrr_at_1 value: 25.358000000000004
  - type: mrr_at_10 value: 38.5
  - type: mrr_at_100 value: 39.572
  - type: mrr_at_1000 value: 39.607
  - type: mrr_at_3 value: 34.491
  - type: mrr_at_5 value: 36.83
  - type: ndcg_at_1 value: 25.358000000000004
  - type: ndcg_at_10 value: 45.214999999999996
  - type: ndcg_at_100 value: 50.56
  - type: ndcg_at_1000 value: 51.507999999999996
  - type: ndcg_at_3 value: 36.925999999999995
  - type: ndcg_at_5 value: 41.182
  - type: precision_at_1 value: 25.358000000000004
  - type: precision_at_10 value: 7.090000000000001
  - type: precision_at_100 value: 0.9740000000000001
  - type: precision_at_1000 value: 0.106
  - type: precision_at_3 value: 15.697
  - type: precision_at_5 value: 11.599
  - type: recall_at_1 value: 24.624
  - type: recall_at_10 value: 67.78699999999999
  - type: recall_at_100 value: 92.11200000000001
  - type: recall_at_1000 value: 99.208
  - type: recall_at_3 value: 45.362
  - type: recall_at_5 value: 55.58
- task: type: Classification dataset: type: mteb/mtop_domain name: MTEB MTOPDomainClassification (en) config: en split: test revision: d80d48c1eb48d3562165c59d59d0034df9fff0bf metrics:
  - type: accuracy value: 96.83310533515733
  - type: f1 value: 96.57069781347995
- task: type: Classification dataset: type: mteb/mtop_intent name: MTEB MTOPIntentClassification (en) config: en split: test revision: ae001d0e6b1228650b7bd1c2c65fb50ad11a8aba metrics:
  - type: accuracy value: 89.5690834473324
  - type: f1 value: 73.7275204564728
- task: type: Classification dataset: type: mteb/amazon_massive_intent name: MTEB MassiveIntentClassification (en) config: en split: test revision: 31efe3c427b0bae9c22cbb560b8f15491cc6bed7 metrics:
  - type: accuracy value: 82.67316745124411
  - type: f1 value: 79.70626515721662
- task: type: Classification dataset: type: mteb/amazon_massive_scenario name: MTEB MassiveScenarioClassification (en) config: en split: test revision: 7d571f92784cd94a019292a1f45445077d0ef634 metrics:
  - type: accuracy value: 85.01344989912575
  - type: f1 value: 84.45181022816965
- task: type: Clustering dataset: type: mteb/medrxiv-clustering-p2p name: MTEB MedrxivClusteringP2P config: default split: test revision: e7a26af6f3ae46b30dde8737f02c07b1505bcc73 metrics:
  - type: v_measure value: 37.843426126777295
- task: type: Clustering dataset: type: mteb/medrxiv-clustering-s2s name: MTEB MedrxivClusteringS2S config: default split: test revision: 35191c8c0dca72d8ff3efcd72aa802307d469663 metrics:
  - type: v_measure value: 36.651728547241476
- task: type: Reranking dataset: type: mteb/mind_small name: MTEB MindSmallReranking config: default split: test revision: 3bdac13927fdc888b903db93b2ffdbd90b295a69 metrics:
  - type: map value: 32.05750522793288
  - type: mrr value: 33.28067556869468
- task: type: Retrieval dataset: type: mteb/nfcorpus name: MTEB NFCorpus config: default split: test revision: ec0fa4fe99da2ff19ca1214b7966684033a58814 metrics:
  - type: map_at_1 value: 6.744
  - type: map_at_10 value: 16.235
  - type: map_at_100 value: 20.767
  - type: map_at_1000 value: 22.469
  - type: map_at_3 value: 11.708
  - type: map_at_5 value: 13.924
  - type: mrr_at_1 value: 55.728
  - type: mrr_at_10 value: 63.869
  - type: mrr_at_100 value: 64.322
  - type: mrr_at_1000 value: 64.342
  - type: mrr_at_3 value: 62.022999999999996
  - type: mrr_at_5 value: 63.105999999999995
  - type: ndcg_at_1 value: 53.096
  - type: ndcg_at_10 value: 41.618
  - type: ndcg_at_100 value: 38.562999999999995
  - type: ndcg_at_1000 value: 47.006
  - type: ndcg_at_3 value: 47.657
  - type: ndcg_at_5 value: 45.562999999999995
  - type: precision_at_1 value: 55.108000000000004
  - type: precision_at_10 value: 30.464000000000002
  - type: precision_at_100 value: 9.737
  - type: precision_at_1000 value: 2.2720000000000002
  - type: precision_at_3 value: 44.376
  - type: precision_at_5 value: 39.505
  - type: recall_at_1 value: 6.744
  - type: recall_at_10 value: 21.11
  - type: recall_at_100 value: 39.69
  - type: recall_at_1000 value: 70.44
  - type: recall_at_3 value: 13.120000000000001
  - type: recall_at_5 value: 16.669
- task: type: Retrieval dataset: type: mteb/nq name: MTEB NQ config: default split: test revision: b774495ed302d8c44a3a7ea25c90dbce03968f31 metrics:
  - type: map_at_1 value: 46.263
  - type: map_at_10 value: 63.525
  - type: map_at_100 value: 64.142
  - type: map_at_1000 value: 64.14800000000001
  - type: map_at_3