%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
EN
Mmlw Roberta Base
M
Mmlw Roberta Base
Developed by sdadas
A Polish sentence embedding model based on RoBERTa architecture, focusing on sentence similarity calculation and feature extraction tasks.
Downloads 106.30k
Release Time : 11/17/2023
Model Overview
This model is a Polish pre-trained model based on RoBERTa architecture, mainly used for sentence similarity calculation, feature extraction, and related natural language processing tasks. It demonstrates multiple capabilities in the MTEB (Massive Text Embedding Benchmark) Polish benchmark tests.
Model Features
Polish language optimization
Specially optimized for Polish text, excelling in Polish NLP tasks.
Versatile sentence embeddings
Capable of generating high-quality sentence embeddings suitable for various downstream tasks.
MTEB benchmark validation
Comprehensively evaluated across multiple MTEB Polish benchmark tasks.
Model Capabilities
Sentence similarity calculation
Text feature extraction
Text clustering
Text classification
Information retrieval
Semantic text similarity evaluation
Use Cases
Text analysis
Review classification
Sentiment or topic classification of Polish product reviews
Achieved 40.26% accuracy on the AllegroReviews dataset
Information retrieval
Document retrieval
Polish document similarity search and retrieval
Achieved 25.24% map@1000 on the DBPedia-PL dataset
Semantic analysis
Semantic similarity evaluation
Evaluating semantic similarity of Polish sentence pairs
Achieved 92.55 Spearman correlation coefficient on the CDSC-R dataset
pipeline_tag: sentence-similarity tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
- mteb model-index:
- name: mmlw-roberta-base
results:
- task:
type: Clustering
dataset:
type: PL-MTEB/8tags-clustering
name: MTEB 8TagsClustering
config: default
split: test
revision: None
metrics:
- type: v_measure value: 33.08463724780795
- task:
type: Classification
dataset:
type: PL-MTEB/allegro-reviews
name: MTEB AllegroReviews
config: default
split: test
revision: None
metrics:
- type: accuracy value: 40.25844930417495
- type: f1 value: 35.59685265418916
- task:
type: Retrieval
dataset:
type: arguana-pl
name: MTEB ArguAna-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 33.073
- type: map_at_10 value: 50.223
- type: map_at_100 value: 50.942
- type: map_at_1000 value: 50.94499999999999
- type: map_at_3 value: 45.721000000000004
- type: map_at_5 value: 48.413000000000004
- type: mrr_at_1 value: 34.424
- type: mrr_at_10 value: 50.68899999999999
- type: mrr_at_100 value: 51.437999999999995
- type: mrr_at_1000 value: 51.441
- type: mrr_at_3 value: 46.219
- type: mrr_at_5 value: 48.921
- type: ndcg_at_1 value: 33.073
- type: ndcg_at_10 value: 59.021
- type: ndcg_at_100 value: 61.902
- type: ndcg_at_1000 value: 61.983999999999995
- type: ndcg_at_3 value: 49.818
- type: ndcg_at_5 value: 54.644999999999996
- type: precision_at_1 value: 33.073
- type: precision_at_10 value: 8.684
- type: precision_at_100 value: 0.9900000000000001
- type: precision_at_1000 value: 0.1
- type: precision_at_3 value: 20.555
- type: precision_at_5 value: 14.666
- type: recall_at_1 value: 33.073
- type: recall_at_10 value: 86.842
- type: recall_at_100 value: 99.004
- type: recall_at_1000 value: 99.644
- type: recall_at_3 value: 61.663999999999994
- type: recall_at_5 value: 73.329
- task:
type: Classification
dataset:
type: PL-MTEB/cbd
name: MTEB CBD
config: default
split: test
revision: None
metrics:
- type: accuracy value: 68.11
- type: ap value: 20.916633959031266
- type: f1 value: 56.85804802205465
- task:
type: PairClassification
dataset:
type: PL-MTEB/cdsce-pairclassification
name: MTEB CDSC-E
config: default
split: test
revision: None
metrics:
- type: cos_sim_accuracy value: 89.2
- type: cos_sim_ap value: 79.1041156765933
- type: cos_sim_f1 value: 70.0
- type: cos_sim_precision value: 74.11764705882354
- type: cos_sim_recall value: 66.3157894736842
- type: dot_accuracy value: 88.2
- type: dot_ap value: 72.57183688228149
- type: dot_f1 value: 67.16417910447761
- type: dot_precision value: 63.67924528301887
- type: dot_recall value: 71.05263157894737
- type: euclidean_accuracy value: 89.3
- type: euclidean_ap value: 79.01345533432428
- type: euclidean_f1 value: 70.19498607242339
- type: euclidean_precision value: 74.55621301775149
- type: euclidean_recall value: 66.3157894736842
- type: manhattan_accuracy value: 89.3
- type: manhattan_ap value: 79.01671381791259
- type: manhattan_f1 value: 70.0280112044818
- type: manhattan_precision value: 74.8502994011976
- type: manhattan_recall value: 65.78947368421053
- type: max_accuracy value: 89.3
- type: max_ap value: 79.1041156765933
- type: max_f1 value: 70.19498607242339
- task:
type: STS
dataset:
type: PL-MTEB/cdscr-sts
name: MTEB CDSC-R
config: default
split: test
revision: None
metrics:
- type: cos_sim_pearson value: 91.79559442663039
- type: cos_sim_spearman value: 92.5438168962641
- type: euclidean_pearson value: 92.02981265332856
- type: euclidean_spearman value: 92.5548245733484
- type: manhattan_pearson value: 91.95296287979178
- type: manhattan_spearman value: 92.50279516120241
- task:
type: Retrieval
dataset:
type: dbpedia-pl
name: MTEB DBPedia-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 7.829999999999999
- type: map_at_10 value: 16.616
- type: map_at_100 value: 23.629
- type: map_at_1000 value: 25.235999999999997
- type: map_at_3 value: 12.485
- type: map_at_5 value: 14.077
- type: mrr_at_1 value: 61.75000000000001
- type: mrr_at_10 value: 69.852
- type: mrr_at_100 value: 70.279
- type: mrr_at_1000 value: 70.294
- type: mrr_at_3 value: 68.375
- type: mrr_at_5 value: 69.187
- type: ndcg_at_1 value: 49.75
- type: ndcg_at_10 value: 36.217
- type: ndcg_at_100 value: 41.235
- type: ndcg_at_1000 value: 48.952
- type: ndcg_at_3 value: 41.669
- type: ndcg_at_5 value: 38.285000000000004
- type: precision_at_1 value: 61.5
- type: precision_at_10 value: 28.499999999999996
- type: precision_at_100 value: 9.572
- type: precision_at_1000 value: 2.025
- type: precision_at_3 value: 44.083
- type: precision_at_5 value: 36.3
- type: recall_at_1 value: 7.829999999999999
- type: recall_at_10 value: 21.462999999999997
- type: recall_at_100 value: 47.095
- type: recall_at_1000 value: 71.883
- type: recall_at_3 value: 13.891
- type: recall_at_5 value: 16.326999999999998
- task:
type: Retrieval
dataset:
type: fiqa-pl
name: MTEB FiQA-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 16.950000000000003
- type: map_at_10 value: 27.422
- type: map_at_100 value: 29.146
- type: map_at_1000 value: 29.328
- type: map_at_3 value: 23.735999999999997
- type: map_at_5 value: 25.671
- type: mrr_at_1 value: 33.796
- type: mrr_at_10 value: 42.689
- type: mrr_at_100 value: 43.522
- type: mrr_at_1000 value: 43.563
- type: mrr_at_3 value: 40.226
- type: mrr_at_5 value: 41.685
- type: ndcg_at_1 value: 33.642
- type: ndcg_at_10 value: 35.008
- type: ndcg_at_100 value: 41.839
- type: ndcg_at_1000 value: 45.035
- type: ndcg_at_3 value: 31.358999999999998
- type: ndcg_at_5 value: 32.377
- type: precision_at_1 value: 33.642
- type: precision_at_10 value: 9.937999999999999
- type: precision_at_100 value: 1.685
- type: precision_at_1000 value: 0.22699999999999998
- type: precision_at_3 value: 21.142
- type: precision_at_5 value: 15.586
- type: recall_at_1 value: 16.950000000000003
- type: recall_at_10 value: 42.286
- type: recall_at_100 value: 68.51899999999999
- type: recall_at_1000 value: 87.471
- type: recall_at_3 value: 28.834
- type: recall_at_5 value: 34.274
- task:
type: Retrieval
dataset:
type: hotpotqa-pl
name: MTEB HotpotQA-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 37.711
- type: map_at_10 value: 57.867999999999995
- type: map_at_100 value: 58.77
- type: map_at_1000 value: 58.836999999999996
- type: map_at_3 value: 54.400999999999996
- type: map_at_5 value: 56.564
- type: mrr_at_1 value: 75.449
- type: mrr_at_10 value: 81.575
- type: mrr_at_100 value: 81.783
- type: mrr_at_1000 value: 81.792
- type: mrr_at_3 value: 80.50399999999999
- type: mrr_at_5 value: 81.172
- type: ndcg_at_1 value: 75.422
- type: ndcg_at_10 value: 66.635
- type: ndcg_at_100 value: 69.85
- type: ndcg_at_1000 value: 71.179
- type: ndcg_at_3 value: 61.648
- type: ndcg_at_5 value: 64.412
- type: precision_at_1 value: 75.422
- type: precision_at_10 value: 13.962
- type: precision_at_100 value: 1.649
- type: precision_at_1000 value: 0.183
- type: precision_at_3 value: 39.172000000000004
- type: precision_at_5 value: 25.691000000000003
- type: recall_at_1 value: 37.711
- type: recall_at_10 value: 69.811
- type: recall_at_100 value: 82.471
- type: recall_at_1000 value: 91.29
- type: recall_at_3 value: 58.757999999999996
- type: recall_at_5 value: 64.227
- task:
type: Retrieval
dataset:
type: msmarco-pl
name: MTEB MSMARCO-PL
config: default
split: validation
revision: None
metrics:
- type: map_at_1 value: 17.033
- type: map_at_10 value: 27.242
- type: map_at_100 value: 28.451999999999998
- type: map_at_1000 value: 28.515
- type: map_at_3 value: 24.046
- type: map_at_5 value: 25.840999999999998
- type: mrr_at_1 value: 17.493
- type: mrr_at_10 value: 27.67
- type: mrr_at_100 value: 28.823999999999998
- type: mrr_at_1000 value: 28.881
- type: mrr_at_3 value: 24.529999999999998
- type: mrr_at_5 value: 26.27
- type: ndcg_at_1 value: 17.479
- type: ndcg_at_10 value: 33.048
- type: ndcg_at_100 value: 39.071
- type: ndcg_at_1000 value: 40.739999999999995
- type: ndcg_at_3 value: 26.493
- type: ndcg_at_5 value: 29.701
- type: precision_at_1 value: 17.479
- type: precision_at_10 value: 5.324
- type: precision_at_100 value: 0.8380000000000001
- type: precision_at_1000 value: 0.098
- type: precision_at_3 value: 11.408999999999999
- type: precision_at_5 value: 8.469999999999999
- type: recall_at_1 value: 17.033
- type: recall_at_10 value: 50.929
- type: recall_at_100 value: 79.262
- type: recall_at_1000 value: 92.239
- type: recall_at_3 value: 33.06
- type: recall_at_5 value: 40.747
- task:
type: Classification
dataset:
type: mteb/amazon_massive_intent
name: MTEB MassiveIntentClassification (pl)
config: pl
split: test
revision: 31efe3c427b0bae9c22cbb560b8f15491cc6bed7
metrics:
- type: accuracy value: 72.31002017484867
- type: f1 value: 69.61603671063031
- task:
type: Classification
dataset:
type: mteb/amazon_massive_scenario
name: MTEB MassiveScenarioClassification (pl)
config: pl
split: test
revision: 7d571f92784cd94a019292a1f45445077d0ef634
metrics:
- type: accuracy value: 75.52790854068594
- type: f1 value: 75.4053872472259
- task:
type: Retrieval
dataset:
type: nfcorpus-pl
name: MTEB NFCorpus-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 5.877000000000001
- type: map_at_10 value: 12.817
- type: map_at_100 value: 16.247
- type: map_at_1000 value: 17.683
- type: map_at_3 value: 9.334000000000001
- type: map_at_5 value: 10.886999999999999
- type: mrr_at_1 value: 45.201
- type: mrr_at_10 value: 52.7
- type: mrr_at_100 value: 53.425999999999995
- type: mrr_at_1000 value: 53.461000000000006
- type: mrr_at_3 value: 50.464
- type: mrr_at_5 value: 51.827
- type: ndcg_at_1 value: 41.949999999999996
- type: ndcg_at_10 value: 34.144999999999996
- type: ndcg_at_100 value: 31.556
- type: ndcg_at_1000 value: 40.265
- type: ndcg_at_3 value: 38.07
- type: ndcg_at_5 value: 36.571
- type: precision_at_1 value: 44.272
- type: precision_at_10 value: 25.697
- type: precision_at_100 value: 8.077
- type: precision_at_1000 value: 2.084
- type: precision_at_3 value: 36.016999999999996
- type: precision_at_5 value: 31.703
- type: recall_at_1 value: 5.877000000000001
- type: recall_at_10 value: 16.986
- type: recall_at_100 value: 32.719
- type: recall_at_1000 value: 63.763000000000005
- type: recall_at_3 value: 10.292
- type: recall_at_5 value: 12.886000000000001
- task:
type: Retrieval
dataset:
type: nq-pl
name: MTEB NQ-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 25.476
- type: map_at_10 value: 38.67
- type: map_at_100 value: 39.784000000000006
- type: map_at_1000 value: 39.831
- type: map_at_3 value: 34.829
- type: map_at_5 value: 37.025000000000006
- type: mrr_at_1 value: 28.621000000000002
- type: mrr_at_10 value: 41.13
- type: mrr_at_100 value: 42.028
- type: mrr_at_1000 value: 42.059999999999995
- type: mrr_at_3 value: 37.877
- type: mrr_at_5 value: 39.763999999999996
- type: ndcg_at_1 value: 28.563
- type: ndcg_at_10 value: 45.654
- type: ndcg_at_100 value: 50.695
- type: ndcg_at_1000 value: 51.873999999999995
- type: ndcg_at_3 value: 38.359
- type: ndcg_at_5 value: 42.045
- type: precision_at_1 value: 28.563
- type: precision_at_10 value: 7.6450000000000005
- type: precision_at_100 value: 1.052
- type: precision_at_1000 value: 0.117
- type: precision_at_3 value: 17.458000000000002
- type: precision_at_5 value: 12.613
- type: recall_at_1 value: 25.476
- type: recall_at_10 value: 64.484
- type: recall_at_100 value: 86.96199999999999
- type: recall_at_1000 value: 95.872
- type: recall_at_3 value: 45.527
- type: recall_at_5 value: 54.029
- task:
type: Classification
dataset:
type: laugustyniak/abusive-clauses-pl
name: MTEB PAC
config: default
split: test
revision: None
metrics:
- type: accuracy value: 65.87315377932232
- type: ap value: 76.41966964416534
- type: f1 value: 63.64417488639012
- task:
type: PairClassification
dataset:
type: PL-MTEB/ppc-pairclassification
name: MTEB PPC
config: default
split: test
revision: None
metrics:
- type: cos_sim_accuracy value: 87.7
- type: cos_sim_ap value: 92.81319372631636
- type: cos_sim_f1 value: 90.04048582995952
- type: cos_sim_precision value: 88.11410459587957
- type: cos_sim_recall value: 92.05298013245033
- type: dot_accuracy value: 75.0
- type: dot_ap value: 83.63089957943261
- type: dot_f1 value: 80.76923076923077
- type: dot_precision value: 75.43103448275862
- type: dot_recall value: 86.9205298013245
- type: euclidean_accuracy value: 87.7
- type: euclidean_ap value: 92.94772245932825
- type: euclidean_f1 value: 90.10458567980692
- type: euclidean_precision value: 87.63693270735524
- type: euclidean_recall value: 92.71523178807946
- type: manhattan_accuracy value: 87.8
- type: manhattan_ap value: 92.95330512127123
- type: manhattan_f1 value: 90.08130081300813
- type: manhattan_precision value: 88.49840255591054
- type: manhattan_recall value: 91.72185430463577
- type: max_accuracy value: 87.8
- type: max_ap value: 92.95330512127123
- type: max_f1 value: 90.10458567980692
- task:
type: PairClassification
dataset:
type: PL-MTEB/psc-pairclassification
name: MTEB PSC
config: default
split: test
revision: None
metrics:
- type: cos_sim_accuracy value: 96.19666048237477
- type: cos_sim_ap value: 98.61237969571302
- type: cos_sim_f1 value: 93.77845220030349
- type: cos_sim_precision value: 93.35347432024169
- type: cos_sim_recall value: 94.20731707317073
- type: dot_accuracy value: 94.89795918367348
- type: dot_ap value: 97.02853491357943
- type: dot_f1 value: 91.85185185185186
- type: dot_precision value: 89.33717579250721
- type: dot_recall value: 94.51219512195121
- type: euclidean_accuracy value: 96.38218923933209
- type: euclidean_ap value: 98.58145584134218
- type: euclidean_f1 value: 94.04580152671755
- type: euclidean_precision value: 94.18960244648318
- type: euclidean_recall value: 93.90243902439023
- type: manhattan_accuracy value: 96.47495361781077
- type: manhattan_ap value: 98.6108221024781
- type: manhattan_f1 value: 94.18960244648318
- type: manhattan_precision value: 94.47852760736197
- type: manhattan_recall value: 93.90243902439023
- type: max_accuracy value: 96.47495361781077
- type: max_ap value: 98.61237969571302
- type: max_f1 value: 94.18960244648318
- task:
type: Classification
dataset:
type: PL-MTEB/polemo2_in
name: MTEB PolEmo2.0-IN
config: default
split: test
revision: None
metrics:
- type: accuracy value: 71.73130193905818
- type: f1 value: 71.17731918813324
- task:
type: Classification
dataset:
type: PL-MTEB/polemo2_out
name: MTEB PolEmo2.0-OUT
config: default
split: test
revision: None
metrics:
- type: accuracy value: 46.59919028340081
- type: f1 value: 37.216392949948954
- task:
type: Retrieval
dataset:
type: quora-pl
name: MTEB Quora-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 66.134
- type: map_at_10 value: 80.19
- type: map_at_100 value: 80.937
- type: map_at_1000 value: 80.95599999999999
- type: map_at_3 value: 77.074
- type: map_at_5 value: 79.054
- type: mrr_at_1 value: 75.88000000000001
- type: mrr_at_10 value: 83.226
- type: mrr_at_100 value: 83.403
- type: mrr_at_1000 value: 83.406
- type: mrr_at_3 value: 82.03200000000001
- type: mrr_at_5 value: 82.843
- type: ndcg_at_1 value: 75.94
- type: ndcg_at_10 value: 84.437
- type: ndcg_at_100 value: 86.13
- type: ndcg_at_1000 value: 86.29299999999999
- type: ndcg_at_3 value: 81.07799999999999
- type: ndcg_at_5 value: 83.0
- type: precision_at_1 value: 75.94
- type: precision_at_10 value: 12.953999999999999
- type: precision_at_100 value: 1.514
- type: precision_at_1000 value: 0.156
- type: precision_at_3 value: 35.61
- type: precision_at_5 value: 23.652
- type: recall_at_1 value: 66.134
- type: recall_at_10 value: 92.991
- type: recall_at_100 value: 99.003
- type: recall_at_1000 value: 99.86
- type: recall_at_3 value: 83.643
- type: recall_at_5 value: 88.81099999999999
- task:
type: Retrieval
dataset:
type: scidocs-pl
name: MTEB SCIDOCS-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 4.183
- type: map_at_10 value: 10.626
- type: map_at_100 value: 12.485
- type: map_at_1000 value: 12.793
- type: map_at_3 value: 7.531000000000001
- type: map_at_5 value: 9.037
- type: mrr_at_1 value: 20.5
- type: mrr_at_10 value: 30.175
- type: mrr_at_100 value: 31.356
- type: mrr_at_1000 value: 31.421
- type: mrr_at_3 value: 26.900000000000002
- type: mrr_at_5 value: 28.689999999999998
- type: ndcg_at_1 value: 20.599999999999998
- type: ndcg_at_10 value: 17.84
- type: ndcg_at_100 value: 25.518
- type: ndcg_at_1000 value: 31.137999999999998
- type: ndcg_at_3 value: 16.677
- type: ndcg_at_5 value: 14.641000000000002
- type: precision_at_1 value: 20.599999999999998
- type: precision_at_10 value: 9.3
- type: precision_at_100 value: 2.048
- type: precision_at_1000 value: 0.33999999999999997
- type: precision_at_3 value: 15.533
- type: precision_at_5 value: 12.839999999999998
- type: recall_at_1 value: 4.183
- type: recall_at_10 value: 18.862000000000002
- type: recall_at_100 value: 41.592
- type: recall_at_1000 value: 69.037
- type: recall_at_3 value: 9.443
- type: recall_at_5 value: 13.028
- task:
type: PairClassification
dataset:
type: PL-MTEB/sicke-pl-pairclassification
name: MTEB SICK-E-PL
config: default
split: test
revision: None
metrics:
- type: cos_sim_accuracy value: 86.32286995515696
- type: cos_sim_ap value: 82.04302619416443
- type: cos_sim_f1 value: 74.95572086432874
- type: cos_sim_precision value: 74.55954897815363
- type: cos_sim_recall value: 75.35612535612536
- type: dot_accuracy value: 83.9176518548716
- type: dot_ap value: 76.8608733580272
- type: dot_f1 value: 72.31936654569449
- type: dot_precision value: 67.36324523663184
- type: dot_recall value: 78.06267806267806
- type: euclidean_accuracy value: 86.32286995515696
- type: euclidean_ap value: 81.9648986659308
- type: euclidean_f1 value: 74.93796526054591
- type: euclidean_precision value: 74.59421312632321
- type: euclidean_recall value: 75.28490028490027
- type: manhattan_accuracy value: 86.30248675091724
- type: manhattan_ap value: 81.92853980116878
- type: manhattan_f1 value: 74.80968858131489
- type: manhattan_precision value: 72.74562584118439
- type: manhattan_recall value: 76.99430199430199
- type: max_accuracy value: 86.32286995515696
- type: max_ap value: 82.04302619416443
- type: max_f1 value: 74.95572086432874
- task:
type: STS
dataset:
type: PL-MTEB/sickr-pl-sts
name: MTEB SICK-R-PL
config: default
split: test
revision: None
metrics:
- type: cos_sim_pearson value: 83.07566183637853
- type: cos_sim_spearman value: 79.20198022242548
- type: euclidean_pearson value: 81.27875473517936
- type: euclidean_spearman value: 79.21560102311153
- type: manhattan_pearson value: 81.21559474880459
- type: manhattan_spearman value: 79.1537846814979
- task:
type: STS
dataset:
type: mteb/sts22-crosslingual-sts
name: MTEB STS22 (pl)
config: pl
split: test
revision: 6d1ba47164174a496b7fa5d3569dae26a6813b80
metrics:
- type: cos_sim_pearson value: 36.39657573900194
- type: cos_sim_spearman value: 40.36403461037013
- type: euclidean_pearson value: 29.143416004776316
- type: euclidean_spearman value: 40.43197841306375
- type: manhattan_pearson value: 29.18632337290767
- type: manhattan_spearman value: 40.50563343395481
- task:
type: Retrieval
dataset:
type: scifact-pl
name: MTEB SciFact-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 49.428
- type: map_at_10 value: 60.423
- type: map_at_100 value: 61.037
- type: map_at_1000 value: 61.065999999999995
- type: map_at_3 value: 56.989000000000004
- type: map_at_5 value: 59.041999999999994
- type: mrr_at_1 value: 52.666999999999994
- type: mrr_at_10 value: 61.746
- type: mrr_at_100 value: 62.273
- type: mrr_at_1000 value: 62.300999999999995
- type: mrr_at_3 value: 59.278
- type: mrr_at_5 value: 60.611000000000004
- type: ndcg_at_1 value: 52.333
- type: ndcg_at_10 value: 65.75
- type: ndcg_at_100 value: 68.566
- type: ndcg_at_1000 value: 69.314
- type: ndcg_at_3 value: 59.768
- type: ndcg_at_5 value: 62.808
- type: precision_at_1 value: 52.333
- type: precision_at_10 value: 9.167
- type: precision_at_100 value: 1.0630000000000002
- type: precision_at_1000 value: 0.11299999999999999
- type: precision_at_3 value: 23.778
- type: precision_at_5 value: 16.2
- type: recall_at_1 value: 49.428
- type: recall_at_10 value: 81.07799999999999
- type: recall_at_100 value: 93.93299999999999
- type: recall_at_1000 value: 99.667
- type: recall_at_3 value: 65.061
- type: recall_at_5 value: 72.667
- task:
type: Retrieval
dataset:
type: trec-covid-pl
name: MTEB TRECCOVID-PL
config: default
split: test
revision: None
metrics:
- type: map_at_1 value: 0.22100000000000003
- type: map_at_10 value: 1.788
- type: map_at_100 value: 9.937
- type: map_at_1000 value: 24.762999999999998
- type: map_at_3 value: 0.579
- type: map_at_5 value: 0.947
- type: mrr_at_1 value: 78.0
- type: mrr_at_10 value: 88.067
- type: mrr_at_100 value: 88.067
- type: mrr_at_1000 value: 88.067
- type: mrr_at_3 value: 87.667
- type: mrr_at_5 value: 88.067
- type: ndcg_at_1 value: 76.0
- type: ndcg_at_10 value: 71.332
- type: ndcg_at_100 value: 54.80500000000001
- type: ndcg_at_1000 value: 49.504999999999995
- type: ndcg_at_3 value: 73.693
- type: ndcg_at_5 value: 73.733
- type: precision_at_1 value: 82.0
- type: precision_at_10 value: 76.8
- type: precision_at_100 value: 56.68
- type: precision_at_1000 value: 22.236
- type: precision_at_3 value: 78.667
- type: precision_at_5 value: 79.2
- type: recall_at_1 value: 0.22100000000000003
- type: recall_at_10 value: 2.033
- type: recall_at_100 value: 13.431999999999999
- type: recall_at_1000 value: 46.913
- type: recall_at_3 value: 0.625
- type: recall_at_5 value: 1.052 language: pl license: apache-2.0 widget:
- task:
type: Clustering
dataset:
type: PL-MTEB/8tags-clustering
name: MTEB 8TagsClustering
config: default
split: test
revision: None
metrics:
- source_sentence: "zapytanie: Jak dożyć 100 lat?"
sentences:
- "Trzeba zdrowo się odżywiać i uprawiać sport."
- "Trzeba pić alkohol, imprezować i jeździć szybkimi autami."
- "Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
MMLW-roberta-base
MMLW (muszę mieć lepszą wiadomość) are neural text encoders for Polish. This is a distilled model that can be used to generate embeddings applicable to many tasks such as semantic similarity, clustering, information retrieval. The model can also serve as a base for further fine-tuning. It transforms texts to 768 dimensional vectors. The model was initialized with Polish RoBERTa checkpoint, and then trained with multilingual knowledge distillation method on a diverse corpus of 60 million Polish-English text pairs. We utilised English FlagEmbeddings (BGE) as teacher models for distillation.
Usage (Sentence-Transformers)
⚠️ Our embedding models require the use of specific prefixes and suffixes when encoding texts. For this model, each query should be preceded by the prefix "zapytanie: " ⚠️
You can use the model like this with sentence-transformers:
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim
query_prefix = "zapytanie: "
answer_prefix = ""
queries = [query_prefix + "Jak dożyć 100 lat?"]
answers = [
answer_prefix + "Trzeba zdrowo się odżywiać i uprawiać sport.",
answer_prefix + "Trzeba pić alkohol, imprezować i jeździć szybkimi autami.",
answer_prefix + "Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
]
model = SentenceTransformer("sdadas/mmlw-roberta-base")
queries_emb = model.encode(queries, convert_to_tensor=True, show_progress_bar=False)
answers_emb = model.encode(answers, convert_to_tensor=True, show_progress_bar=False)
best_answer = cos_sim(queries_emb, answers_emb).argmax().item()
print(answers[best_answer])
# Trzeba zdrowo się odżywiać i uprawiać sport.
Evaluation Results
- The model achieves an Average Score of 61.05 on the Polish Massive Text Embedding Benchmark (MTEB). See MTEB Leaderboard for detailed results.
- The model achieves NDCG@10 of 53.60 on the Polish Information Retrieval Benchmark. See PIRB Leaderboard for detailed results.
Acknowledgements
This model was trained with the A100 GPU cluster support delivered by the Gdansk University of Technology within the TASK center initiative.
Citation
@article{dadas2024pirb,
title={{PIRB}: A Comprehensive Benchmark of Polish Dense and Hybrid Text Retrieval Methods},
author={Sławomir Dadas and Michał Perełkiewicz and Rafał Poświata},
year={2024},
eprint={2402.13350},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
Jina Embeddings V3
Jina Embeddings V3 is a multilingual sentence embedding model supporting over 100 languages, specializing in sentence similarity and feature extraction tasks.
Text Embedding
Transformers Supports Multiple Languages
Transformers Supports Multiple LanguagesJ
jinaai
3.7M
911
Ms Marco MiniLM L6 V2
Apache-2.0
A cross-encoder model trained on the MS Marco passage ranking task for query-passage relevance scoring in information retrieval
Text Embedding English
M
cross-encoder
2.5M
86
Opensearch Neural Sparse Encoding Doc V2 Distill
Apache-2.0
A sparse retrieval model based on distillation technology, optimized for OpenSearch, supporting inference-free document encoding with improved search relevance and efficiency over V1
Text Embedding Transformers English
Transformers EnglishO
opensearch-project
1.8M
7
Sapbert From PubMedBERT Fulltext
Apache-2.0
A biomedical entity representation model based on PubMedBERT, optimized for semantic relation capture through self-aligned pre-training
Text Embedding English
S
cambridgeltl
1.7M
49
Gte Large
MIT
GTE-Large is a powerful sentence transformer model focused on sentence similarity and text embedding tasks, excelling in multiple benchmark tests.
Text Embedding English
G
thenlper
1.5M
278
Gte Base En V1.5
Apache-2.0
GTE-base-en-v1.5 is an English sentence transformer model focused on sentence similarity tasks, excelling in multiple text embedding benchmarks.
Text Embedding Transformers Supports Multiple Languages
Transformers Supports Multiple LanguagesG
Alibaba-NLP
1.5M
63
Gte Multilingual Base
Apache-2.0
GTE Multilingual Base is a multilingual sentence embedding model supporting over 50 languages, suitable for tasks like sentence similarity calculation.
Text Embedding Transformers Supports Multiple Languages
Transformers Supports Multiple LanguagesG
Alibaba-NLP
1.2M
246
Polybert
polyBERT is a chemical language model designed to achieve fully machine-driven ultrafast polymer informatics. It maps PSMILES strings into 600-dimensional dense fingerprints to numerically represent polymer chemical structures.
Text Embedding Transformers
TransformersP
kuelumbus
1.0M
5
Bert Base Turkish Cased Mean Nli Stsb Tr
Apache-2.0
A sentence embedding model based on Turkish BERT, optimized for semantic similarity tasks
Text Embedding Transformers Other
Transformers OtherB
emrecan
1.0M
40
GIST Small Embedding V0
MIT
A text embedding model fine-tuned based on BAAI/bge-small-en-v1.5, trained with the MEDI dataset and MTEB classification task datasets, optimized for query encoding in retrieval tasks.
Text Embedding
Safetensors English
Safetensors EnglishG
avsolatorio
945.68k
29
Featured Recommended AI Models