multilingual-e5-small-pruned

This model is a token-embedding pruned version of intfloat/multilingual-e5-small.

Token-embedding pruning clusters semantically similar tokens in the embedding space (using DBSCAN) and merges each cluster into a single shared embedding, shrinking the vocabulary and reducing memory without retraining the transformer layers.

How to use

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("jangedoo/multilingual-e5-small-pruned", trust_remote_code=True)
embeddings = model.encode(["Hello world", "How are you?"])

Note: trust_remote_code=True is required because the model ships a small custom tokenizer class (pruned_tokenizer.py) that applies the id remapping after tokenization. No additional package installation is needed.

Pruning statistics

	Base	Pruned	Reduction
Vocab size	250,037	172,569	30.98%
Total parameters	117,653,760	87,906,048	25.28%
Embedding parameters	96,014,208	66,266,496	30.98%
Embedding size (MB)	366.3	252.8	113.5 MB saved

Evaluation

Dataset / Metric	Base	Pruned	Relative (base = 1.0)
stsb / stsb_pearson_cosine	0.8092	0.7925	0.9794
stsb / stsb_spearman_cosine	0.8359	0.8014	0.9588
nanobeir / NanoClimateFEVER_cosine_accuracy@1	0.3000	0.2600	0.8667
nanobeir / NanoClimateFEVER_cosine_accuracy@3	0.4200	0.3600	0.8571
nanobeir / NanoClimateFEVER_cosine_accuracy@5	0.5000	0.3800	0.7600
nanobeir / NanoClimateFEVER_cosine_accuracy@10	0.6600	0.5400	0.8182
nanobeir / NanoClimateFEVER_cosine_precision@1	0.3000	0.2600	0.8667
nanobeir / NanoClimateFEVER_cosine_precision@3	0.1533	0.1333	0.8696
nanobeir / NanoClimateFEVER_cosine_precision@5	0.1160	0.0880	0.7586
nanobeir / NanoClimateFEVER_cosine_precision@10	0.0880	0.0680	0.7727
nanobeir / NanoClimateFEVER_cosine_recall@1	0.1500	0.1283	0.8556
nanobeir / NanoClimateFEVER_cosine_recall@3	0.2000	0.1717	0.8583
nanobeir / NanoClimateFEVER_cosine_recall@5	0.2433	0.1817	0.7466
nanobeir / NanoClimateFEVER_cosine_recall@10	0.3530	0.2667	0.7554
nanobeir / NanoClimateFEVER_cosine_ndcg@10	0.2927	0.2364	0.8076
nanobeir / NanoClimateFEVER_cosine_mrr@10	0.3906	0.3305	0.8464
nanobeir / NanoClimateFEVER_cosine_map@100	0.2358	0.1934	0.8202
nanobeir / NanoDBPedia_cosine_accuracy@1	0.5800	0.6400	1.1034
nanobeir / NanoDBPedia_cosine_accuracy@3	0.8400	0.7800	0.9286
nanobeir / NanoDBPedia_cosine_accuracy@5	0.8800	0.8400	0.9545
nanobeir / NanoDBPedia_cosine_accuracy@10	0.9600	0.9200	0.9583
nanobeir / NanoDBPedia_cosine_precision@1	0.5800	0.6400	1.1034
nanobeir / NanoDBPedia_cosine_precision@3	0.5400	0.5200	0.9630
nanobeir / NanoDBPedia_cosine_precision@5	0.5200	0.4920	0.9462
nanobeir / NanoDBPedia_cosine_precision@10	0.4300	0.3980	0.9256
nanobeir / NanoDBPedia_cosine_recall@1	0.0755	0.0895	1.1861
nanobeir / NanoDBPedia_cosine_recall@3	0.1534	0.1405	0.9156
nanobeir / NanoDBPedia_cosine_recall@5	0.2049	0.1976	0.9641
nanobeir / NanoDBPedia_cosine_recall@10	0.3126	0.2802	0.8963
nanobeir / NanoDBPedia_cosine_ndcg@10	0.5371	0.5108	0.9510
nanobeir / NanoDBPedia_cosine_mrr@10	0.7175	0.7211	1.0050
nanobeir / NanoDBPedia_cosine_map@100	0.3988	0.3692	0.9256
nanobeir / NanoFEVER_cosine_accuracy@1	0.6200	0.5400	0.8710
nanobeir / NanoFEVER_cosine_accuracy@3	0.8800	0.8200	0.9318
nanobeir / NanoFEVER_cosine_accuracy@5	0.9400	0.8800	0.9362
nanobeir / NanoFEVER_cosine_accuracy@10	0.9800	0.9600	0.9796
nanobeir / NanoFEVER_cosine_precision@1	0.6200	0.5400	0.8710
nanobeir / NanoFEVER_cosine_precision@3	0.3000	0.2800	0.9333
nanobeir / NanoFEVER_cosine_precision@5	0.1960	0.1800	0.9184
nanobeir / NanoFEVER_cosine_precision@10	0.1020	0.1000	0.9804
nanobeir / NanoFEVER_cosine_recall@1	0.5867	0.5067	0.8636
nanobeir / NanoFEVER_cosine_recall@3	0.8433	0.7967	0.9447
nanobeir / NanoFEVER_cosine_recall@5	0.9033	0.8567	0.9483
nanobeir / NanoFEVER_cosine_recall@10	0.9333	0.9233	0.9893
nanobeir / NanoFEVER_cosine_ndcg@10	0.7897	0.7353	0.9310
nanobeir / NanoFEVER_cosine_mrr@10	0.7592	0.6909	0.9100
nanobeir / NanoFEVER_cosine_map@100	0.7338	0.6652	0.9066
nanobeir / NanoFiQA2018_cosine_accuracy@1	0.3600	0.3200	0.8889
nanobeir / NanoFiQA2018_cosine_accuracy@3	0.5600	0.5200	0.9286
nanobeir / NanoFiQA2018_cosine_accuracy@5	0.6200	0.5800	0.9355
nanobeir / NanoFiQA2018_cosine_accuracy@10	0.6600	0.6800	1.0303
nanobeir / NanoFiQA2018_cosine_precision@1	0.3600	0.3200	0.8889
nanobeir / NanoFiQA2018_cosine_precision@3	0.2400	0.1933	0.8056
nanobeir / NanoFiQA2018_cosine_precision@5	0.1800	0.1560	0.8667
nanobeir / NanoFiQA2018_cosine_precision@10	0.1060	0.0960	0.9057
nanobeir / NanoFiQA2018_cosine_recall@1	0.1801	0.1687	0.9371
nanobeir / NanoFiQA2018_cosine_recall@3	0.3545	0.3174	0.8954
nanobeir / NanoFiQA2018_cosine_recall@5	0.4403	0.3816	0.8666
nanobeir / NanoFiQA2018_cosine_recall@10	0.4878	0.4738	0.9713
nanobeir / NanoFiQA2018_cosine_ndcg@10	0.3956	0.3655	0.9240
nanobeir / NanoFiQA2018_cosine_mrr@10	0.4630	0.4302	0.9292
nanobeir / NanoFiQA2018_cosine_map@100	0.3380	0.2928	0.8664
nanobeir / NanoHotpotQA_cosine_accuracy@1	0.7800	0.6800	0.8718
nanobeir / NanoHotpotQA_cosine_accuracy@3	0.9200	0.9000	0.9783
nanobeir / NanoHotpotQA_cosine_accuracy@5	0.9600	0.9200	0.9583
nanobeir / NanoHotpotQA_cosine_accuracy@10	0.9800	0.9400	0.9592
nanobeir / NanoHotpotQA_cosine_precision@1	0.7800	0.6800	0.8718
nanobeir / NanoHotpotQA_cosine_precision@3	0.5000	0.4533	0.9067
nanobeir / NanoHotpotQA_cosine_precision@5	0.3240	0.3040	0.9383
nanobeir / NanoHotpotQA_cosine_precision@10	0.1720	0.1600	0.9302
nanobeir / NanoHotpotQA_cosine_recall@1	0.3900	0.3400	0.8718
nanobeir / NanoHotpotQA_cosine_recall@3	0.7500	0.6800	0.9067
nanobeir / NanoHotpotQA_cosine_recall@5	0.8100	0.7600	0.9383
nanobeir / NanoHotpotQA_cosine_recall@10	0.8600	0.8000	0.9302
nanobeir / NanoHotpotQA_cosine_ndcg@10	0.7997	0.7254	0.9072
nanobeir / NanoHotpotQA_cosine_mrr@10	0.8600	0.7879	0.9161
nanobeir / NanoHotpotQA_cosine_map@100	0.7435	0.6629	0.8916
nanobeir / NanoMSMARCO_cosine_accuracy@1	0.4200	0.4200	1.0000
nanobeir / NanoMSMARCO_cosine_accuracy@3	0.5800	0.6000	1.0345
nanobeir / NanoMSMARCO_cosine_accuracy@5	0.7600	0.6800	0.8947
nanobeir / NanoMSMARCO_cosine_accuracy@10	0.8600	0.7800	0.9070
nanobeir / NanoMSMARCO_cosine_precision@1	0.4200	0.4200	1.0000
nanobeir / NanoMSMARCO_cosine_precision@3	0.1933	0.2000	1.0345
nanobeir / NanoMSMARCO_cosine_precision@5	0.1520	0.1360	0.8947
nanobeir / NanoMSMARCO_cosine_precision@10	0.0860	0.0780	0.9070
nanobeir / NanoMSMARCO_cosine_recall@1	0.4200	0.4200	1.0000
nanobeir / NanoMSMARCO_cosine_recall@3	0.5800	0.6000	1.0345
nanobeir / NanoMSMARCO_cosine_recall@5	0.7600	0.6800	0.8947
nanobeir / NanoMSMARCO_cosine_recall@10	0.8600	0.7800	0.9070
nanobeir / NanoMSMARCO_cosine_ndcg@10	0.6187	0.5920	0.9568
nanobeir / NanoMSMARCO_cosine_mrr@10	0.5436	0.5332	0.9808
nanobeir / NanoMSMARCO_cosine_map@100	0.5517	0.5444	0.9868
nanobeir / NanoNFCorpus_cosine_accuracy@1	0.4200	0.4000	0.9524
nanobeir / NanoNFCorpus_cosine_accuracy@3	0.5000	0.5000	1.0000
nanobeir / NanoNFCorpus_cosine_accuracy@5	0.5600	0.5600	1.0000
nanobeir / NanoNFCorpus_cosine_accuracy@10	0.6400	0.6000	0.9375
nanobeir / NanoNFCorpus_cosine_precision@1	0.4200	0.4000	0.9524
nanobeir / NanoNFCorpus_cosine_precision@3	0.3267	0.3333	1.0204
nanobeir / NanoNFCorpus_cosine_precision@5	0.3280	0.3040	0.9268
nanobeir / NanoNFCorpus_cosine_precision@10	0.2520	0.2460	0.9762
nanobeir / NanoNFCorpus_cosine_recall@1	0.0148	0.0233	1.5744
nanobeir / NanoNFCorpus_cosine_recall@3	0.0442	0.0428	0.9684
nanobeir / NanoNFCorpus_cosine_recall@5	0.0772	0.0685	0.8880
nanobeir / NanoNFCorpus_cosine_recall@10	0.0999	0.0938	0.9389
nanobeir / NanoNFCorpus_cosine_ndcg@10	0.2937	0.2873	0.9783
nanobeir / NanoNFCorpus_cosine_mrr@10	0.4829	0.4625	0.9577
nanobeir / NanoNFCorpus_cosine_map@100	0.1046	0.1013	0.9678
nanobeir / NanoNQ_cosine_accuracy@1	0.5400	0.3400	0.6296
nanobeir / NanoNQ_cosine_accuracy@3	0.6400	0.5200	0.8125
nanobeir / NanoNQ_cosine_accuracy@5	0.7000	0.6000	0.8571
nanobeir / NanoNQ_cosine_accuracy@10	0.8200	0.7200	0.8780
nanobeir / NanoNQ_cosine_precision@1	0.5400	0.3400	0.6296
nanobeir / NanoNQ_cosine_precision@3	0.2133	0.1733	0.8125
nanobeir / NanoNQ_cosine_precision@5	0.1480	0.1240	0.8378
nanobeir / NanoNQ_cosine_precision@10	0.0900	0.0760	0.8444
nanobeir / NanoNQ_cosine_recall@1	0.4900	0.3400	0.6939
nanobeir / NanoNQ_cosine_recall@3	0.5900	0.5000	0.8475
nanobeir / NanoNQ_cosine_recall@5	0.6700	0.5900	0.8806
nanobeir / NanoNQ_cosine_recall@10	0.8000	0.7000	0.8750
nanobeir / NanoNQ_cosine_ndcg@10	0.6371	0.5086	0.7983
nanobeir / NanoNQ_cosine_mrr@10	0.6107	0.4496	0.7362
nanobeir / NanoNQ_cosine_map@100	0.5816	0.4546	0.7816
nanobeir / NanoQuoraRetrieval_cosine_accuracy@1	0.8800	0.8400	0.9545
nanobeir / NanoQuoraRetrieval_cosine_accuracy@3	1.0000	0.9600	0.9600
nanobeir / NanoQuoraRetrieval_cosine_accuracy@5	1.0000	0.9600	0.9600
nanobeir / NanoQuoraRetrieval_cosine_accuracy@10	1.0000	0.9600	0.9600
nanobeir / NanoQuoraRetrieval_cosine_precision@1	0.8800	0.8400	0.9545
nanobeir / NanoQuoraRetrieval_cosine_precision@3	0.4067	0.3733	0.9180
nanobeir / NanoQuoraRetrieval_cosine_precision@5	0.2520	0.2280	0.9048
nanobeir / NanoQuoraRetrieval_cosine_precision@10	0.1320	0.1180	0.8939
nanobeir / NanoQuoraRetrieval_cosine_recall@1	0.7807	0.7540	0.9658
nanobeir / NanoQuoraRetrieval_cosine_recall@3	0.9587	0.9253	0.9652
nanobeir / NanoQuoraRetrieval_cosine_recall@5	0.9693	0.9320	0.9615
nanobeir / NanoQuoraRetrieval_cosine_recall@10	0.9833	0.9393	0.9553
nanobeir / NanoQuoraRetrieval_cosine_ndcg@10	0.9359	0.8947	0.9560
nanobeir / NanoQuoraRetrieval_cosine_mrr@10	0.9333	0.8967	0.9607
nanobeir / NanoQuoraRetrieval_cosine_map@100	0.9123	0.8732	0.9572
nanobeir / NanoSCIDOCS_cosine_accuracy@1	0.4000	0.3000	0.7500
nanobeir / NanoSCIDOCS_cosine_accuracy@3	0.6400	0.5200	0.8125
nanobeir / NanoSCIDOCS_cosine_accuracy@5	0.7400	0.6000	0.8108
nanobeir / NanoSCIDOCS_cosine_accuracy@10	0.8200	0.7800	0.9512
nanobeir / NanoSCIDOCS_cosine_precision@1	0.4000	0.3000	0.7500
nanobeir / NanoSCIDOCS_cosine_precision@3	0.3067	0.2333	0.7609
nanobeir / NanoSCIDOCS_cosine_precision@5	0.2600	0.2000	0.7692
nanobeir / NanoSCIDOCS_cosine_precision@10	0.1560	0.1400	0.8974
nanobeir / NanoSCIDOCS_cosine_recall@1	0.0847	0.0627	0.7402
nanobeir / NanoSCIDOCS_cosine_recall@3	0.1897	0.1437	0.7575
nanobeir / NanoSCIDOCS_cosine_recall@5	0.2667	0.2057	0.7712
nanobeir / NanoSCIDOCS_cosine_recall@10	0.3187	0.2887	0.9059
nanobeir / NanoSCIDOCS_cosine_ndcg@10	0.3225	0.2703	0.8380
nanobeir / NanoSCIDOCS_cosine_mrr@10	0.5353	0.4398	0.8216
nanobeir / NanoSCIDOCS_cosine_map@100	0.2448	0.1997	0.8155
nanobeir / NanoArguAna_cosine_accuracy@1	0.1000	0.1000	1.0000
nanobeir / NanoArguAna_cosine_accuracy@3	0.4800	0.4400	0.9167
nanobeir / NanoArguAna_cosine_accuracy@5	0.6200	0.4800	0.7742
nanobeir / NanoArguAna_cosine_accuracy@10	0.7200	0.6200	0.8611
nanobeir / NanoArguAna_cosine_precision@1	0.1000	0.1000	1.0000
nanobeir / NanoArguAna_cosine_precision@3	0.1600	0.1467	0.9167
nanobeir / NanoArguAna_cosine_precision@5	0.1240	0.0960	0.7742
nanobeir / NanoArguAna_cosine_precision@10	0.0720	0.0620	0.8611
nanobeir / NanoArguAna_cosine_recall@1	0.1000	0.1000	1.0000
nanobeir / NanoArguAna_cosine_recall@3	0.4800	0.4400	0.9167
nanobeir / NanoArguAna_cosine_recall@5	0.6200	0.4800	0.7742
nanobeir / NanoArguAna_cosine_recall@10	0.7200	0.6200	0.8611
nanobeir / NanoArguAna_cosine_ndcg@10	0.4121	0.3676	0.8920
nanobeir / NanoArguAna_cosine_mrr@10	0.3128	0.2864	0.9156
nanobeir / NanoArguAna_cosine_map@100	0.3267	0.2962	0.9067
nanobeir / NanoSciFact_cosine_accuracy@1	0.6800	0.5200	0.7647
nanobeir / NanoSciFact_cosine_accuracy@3	0.7400	0.6800	0.9189
nanobeir / NanoSciFact_cosine_accuracy@5	0.7400	0.7400	1.0000
nanobeir / NanoSciFact_cosine_accuracy@10	0.7800	0.7800	1.0000
nanobeir / NanoSciFact_cosine_precision@1	0.6800	0.5200	0.7647
nanobeir / NanoSciFact_cosine_precision@3	0.2533	0.2400	0.9474
nanobeir / NanoSciFact_cosine_precision@5	0.1600	0.1600	1.0000
nanobeir / NanoSciFact_cosine_precision@10	0.0880	0.0860	0.9773
nanobeir / NanoSciFact_cosine_recall@1	0.6450	0.5000	0.7752
nanobeir / NanoSciFact_cosine_recall@3	0.7150	0.6600	0.9231
nanobeir / NanoSciFact_cosine_recall@5	0.7250	0.7250	1.0000
nanobeir / NanoSciFact_cosine_recall@10	0.7800	0.7700	0.9872
nanobeir / NanoSciFact_cosine_ndcg@10	0.7209	0.6455	0.8955
nanobeir / NanoSciFact_cosine_mrr@10	0.7117	0.6116	0.8592
nanobeir / NanoSciFact_cosine_map@100	0.7011	0.6058	0.8640
nanobeir / NanoTouche2020_cosine_accuracy@1	0.4898	0.4082	0.8333
nanobeir / NanoTouche2020_cosine_accuracy@3	0.8980	0.7959	0.8864
nanobeir / NanoTouche2020_cosine_accuracy@5	0.9388	0.8776	0.9348
nanobeir / NanoTouche2020_cosine_accuracy@10	0.9796	0.9592	0.9792
nanobeir / NanoTouche2020_cosine_precision@1	0.4898	0.4082	0.8333
nanobeir / NanoTouche2020_cosine_precision@3	0.5442	0.4150	0.7625
nanobeir / NanoTouche2020_cosine_precision@5	0.4816	0.4163	0.8644
nanobeir / NanoTouche2020_cosine_precision@10	0.4000	0.3429	0.8571
nanobeir / NanoTouche2020_cosine_recall@1	0.0309	0.0251	0.8101
nanobeir / NanoTouche2020_cosine_recall@3	0.1093	0.0844	0.7719
nanobeir / NanoTouche2020_cosine_recall@5	0.1638	0.1409	0.8598
nanobeir / NanoTouche2020_cosine_recall@10	0.2602	0.2243	0.8621
nanobeir / NanoTouche2020_cosine_ndcg@10	0.4483	0.3753	0.8372
nanobeir / NanoTouche2020_cosine_mrr@10	0.6885	0.5975	0.8679
nanobeir / NanoTouche2020_cosine_map@100	0.3263	0.2599	0.7967
nanobeir / NanoBEIR_mean_cosine_accuracy@1	0.5054	0.4437	0.8780
nanobeir / NanoBEIR_mean_cosine_accuracy@3	0.6998	0.6458	0.9228
nanobeir / NanoBEIR_mean_cosine_accuracy@5	0.7661	0.6998	0.9135
nanobeir / NanoBEIR_mean_cosine_accuracy@10	0.8354	0.7876	0.9429
nanobeir / NanoBEIR_mean_cosine_precision@1	0.5054	0.4437	0.8780
nanobeir / NanoBEIR_mean_cosine_precision@3	0.3183	0.2842	0.8930
nanobeir / NanoBEIR_mean_cosine_precision@5	0.2494	0.2219	0.8898
nanobeir / NanoBEIR_mean_cosine_precision@10	0.1672	0.1516	0.9066
nanobeir / NanoBEIR_mean_cosine_recall@1	0.3037	0.2660	0.8759
nanobeir / NanoBEIR_mean_cosine_recall@3	0.4591	0.4233	0.9220
nanobeir / NanoBEIR_mean_cosine_recall@5	0.5272	0.4769	0.9045
nanobeir / NanoBEIR_mean_cosine_recall@10	0.5976	0.5508	0.9216
nanobeir / NanoBEIR_mean_cosine_ndcg@10	0.5542	0.5011	0.9043
nanobeir / NanoBEIR_mean_cosine_mrr@10	0.6161	0.5568	0.9037
nanobeir / NanoBEIR_mean_cosine_map@100	0.4769	0.4245	0.8902
nanobeir_ne / NanoClimateFEVER_cosine_accuracy@1	0.1000	0.0600	0.6000
nanobeir_ne / NanoClimateFEVER_cosine_accuracy@3	0.3000	0.1000	0.3333
nanobeir_ne / NanoClimateFEVER_cosine_accuracy@5	0.4200	0.2000	0.4762
nanobeir_ne / NanoClimateFEVER_cosine_accuracy@10	0.5400	0.3400	0.6296
nanobeir_ne / NanoClimateFEVER_cosine_precision@1	0.1000	0.0600	0.6000
nanobeir_ne / NanoClimateFEVER_cosine_precision@3	0.1000	0.0333	0.3333
nanobeir_ne / NanoClimateFEVER_cosine_precision@5	0.1000	0.0400	0.4000
nanobeir_ne / NanoClimateFEVER_cosine_precision@10	0.0700	0.0380	0.5429
nanobeir_ne / NanoClimateFEVER_cosine_recall@1	0.0300	0.0350	1.1667
nanobeir_ne / NanoClimateFEVER_cosine_recall@3	0.1400	0.0450	0.3214
nanobeir_ne / NanoClimateFEVER_cosine_recall@5	0.2073	0.0967	0.4662
nanobeir_ne / NanoClimateFEVER_cosine_recall@10	0.2747	0.1533	0.5583
nanobeir_ne / NanoClimateFEVER_cosine_ndcg@10	0.1836	0.0970	0.5282
nanobeir_ne / NanoClimateFEVER_cosine_mrr@10	0.2279	0.1112	0.4881
nanobeir_ne / NanoClimateFEVER_cosine_map@100	0.1241	0.0700	0.5636
nanobeir_ne / NanoDBPedia_cosine_accuracy@1	0.4000	0.3800	0.9500
nanobeir_ne / NanoDBPedia_cosine_accuracy@3	0.7600	0.6200	0.8158
nanobeir_ne / NanoDBPedia_cosine_accuracy@5	0.8000	0.7400	0.9250
nanobeir_ne / NanoDBPedia_cosine_accuracy@10	0.8200	0.8400	1.0244
nanobeir_ne / NanoDBPedia_cosine_precision@1	0.4000	0.3800	0.9500
nanobeir_ne / NanoDBPedia_cosine_precision@3	0.4133	0.3333	0.8065
nanobeir_ne / NanoDBPedia_cosine_precision@5	0.3680	0.3360	0.9130
nanobeir_ne / NanoDBPedia_cosine_precision@10	0.3260	0.2860	0.8773
nanobeir_ne / NanoDBPedia_cosine_recall@1	0.0736	0.0736	1.0002
nanobeir_ne / NanoDBPedia_cosine_recall@3	0.1475	0.1168	0.7921
nanobeir_ne / NanoDBPedia_cosine_recall@5	0.1746	0.1629	0.9329
nanobeir_ne / NanoDBPedia_cosine_recall@10	0.2453	0.2255	0.9193
nanobeir_ne / NanoDBPedia_cosine_ndcg@10	0.4156	0.3676	0.8845
nanobeir_ne / NanoDBPedia_cosine_mrr@10	0.5748	0.5297	0.9215
nanobeir_ne / NanoDBPedia_cosine_map@100	0.3034	0.2661	0.8770
nanobeir_ne / NanoFEVER_cosine_accuracy@1	0.3400	0.1800	0.5294
nanobeir_ne / NanoFEVER_cosine_accuracy@3	0.5800	0.4600	0.7931
nanobeir_ne / NanoFEVER_cosine_accuracy@5	0.6600	0.5800	0.8788
nanobeir_ne / NanoFEVER_cosine_accuracy@10	0.8000	0.7000	0.8750
nanobeir_ne / NanoFEVER_cosine_precision@1	0.3400	0.1800	0.5294
nanobeir_ne / NanoFEVER_cosine_precision@3	0.1933	0.1533	0.7931
nanobeir_ne / NanoFEVER_cosine_precision@5	0.1360	0.1200	0.8824
nanobeir_ne / NanoFEVER_cosine_precision@10	0.0820	0.0720	0.8780
nanobeir_ne / NanoFEVER_cosine_recall@1	0.3267	0.1800	0.5510
nanobeir_ne / NanoFEVER_cosine_recall@3	0.5567	0.4500	0.8084
nanobeir_ne / NanoFEVER_cosine_recall@5	0.6467	0.5567	0.8608
nanobeir_ne / NanoFEVER_cosine_recall@10	0.7767	0.6767	0.8712
nanobeir_ne / NanoFEVER_cosine_ndcg@10	0.5473	0.4206	0.7685
nanobeir_ne / NanoFEVER_cosine_mrr@10	0.4854	0.3419	0.7043
nanobeir_ne / NanoFEVER_cosine_map@100	0.4794	0.3463	0.7223
nanobeir_ne / NanoFiQA2018_cosine_accuracy@1	0.2600	0.1000	0.3846
nanobeir_ne / NanoFiQA2018_cosine_accuracy@3	0.4200	0.2400	0.5714
nanobeir_ne / NanoFiQA2018_cosine_accuracy@5	0.4600	0.2600	0.5652
nanobeir_ne / NanoFiQA2018_cosine_accuracy@10	0.5400	0.3600	0.6667
nanobeir_ne / NanoFiQA2018_cosine_precision@1	0.2600	0.1000	0.3846
nanobeir_ne / NanoFiQA2018_cosine_precision@3	0.1600	0.0867	0.5417
nanobeir_ne / NanoFiQA2018_cosine_precision@5	0.1240	0.0600	0.4839
nanobeir_ne / NanoFiQA2018_cosine_precision@10	0.0800	0.0400	0.5000
nanobeir_ne / NanoFiQA2018_cosine_recall@1	0.1287	0.0640	0.4971
nanobeir_ne / NanoFiQA2018_cosine_recall@3	0.2288	0.1807	0.7897
nanobeir_ne / NanoFiQA2018_cosine_recall@5	0.2893	0.1872	0.6470
nanobeir_ne / NanoFiQA2018_cosine_recall@10	0.3780	0.2352	0.6221
nanobeir_ne / NanoFiQA2018_cosine_ndcg@10	0.2912	0.1687	0.5793
nanobeir_ne / NanoFiQA2018_cosine_mrr@10	0.3572	0.1806	0.5056
nanobeir_ne / NanoFiQA2018_cosine_map@100	0.2275	0.1385	0.6088
nanobeir_ne / NanoHotpotQA_cosine_accuracy@1	0.7800	0.6600	0.8462
nanobeir_ne / NanoHotpotQA_cosine_accuracy@3	0.8400	0.8000	0.9524
nanobeir_ne / NanoHotpotQA_cosine_accuracy@5	0.8600	0.8200	0.9535
nanobeir_ne / NanoHotpotQA_cosine_accuracy@10	0.9000	0.8400	0.9333
nanobeir_ne / NanoHotpotQA_cosine_precision@1	0.7800	0.6600	0.8462
nanobeir_ne / NanoHotpotQA_cosine_precision@3	0.3800	0.3467	0.9123
nanobeir_ne / NanoHotpotQA_cosine_precision@5	0.2520	0.2200	0.8730
nanobeir_ne / NanoHotpotQA_cosine_precision@10	0.1380	0.1180	0.8551
nanobeir_ne / NanoHotpotQA_cosine_recall@1	0.3900	0.3300	0.8462
nanobeir_ne / NanoHotpotQA_cosine_recall@3	0.5700	0.5200	0.9123
nanobeir_ne / NanoHotpotQA_cosine_recall@5	0.6300	0.5500	0.8730
nanobeir_ne / NanoHotpotQA_cosine_recall@10	0.6900	0.5900	0.8551
nanobeir_ne / NanoHotpotQA_cosine_ndcg@10	0.6636	0.5728	0.8631
nanobeir_ne / NanoHotpotQA_cosine_mrr@10	0.8132	0.7269	0.8938
nanobeir_ne / NanoHotpotQA_cosine_map@100	0.5941	0.5034	0.8473
nanobeir_ne / NanoMSMARCO_cosine_accuracy@1	0.2600	0.1800	0.6923
nanobeir_ne / NanoMSMARCO_cosine_accuracy@3	0.5800	0.4400	0.7586
nanobeir_ne / NanoMSMARCO_cosine_accuracy@5	0.6600	0.5200	0.7879
nanobeir_ne / NanoMSMARCO_cosine_accuracy@10	0.7400	0.6800	0.9189
nanobeir_ne / NanoMSMARCO_cosine_precision@1	0.2600	0.1800	0.6923
nanobeir_ne / NanoMSMARCO_cosine_precision@3	0.1933	0.1467	0.7586
nanobeir_ne / NanoMSMARCO_cosine_precision@5	0.1320	0.1040	0.7879
nanobeir_ne / NanoMSMARCO_cosine_precision@10	0.0740	0.0680	0.9189
nanobeir_ne / NanoMSMARCO_cosine_recall@1	0.2600	0.1800	0.6923
nanobeir_ne / NanoMSMARCO_cosine_recall@3	0.5800	0.4400	0.7586
nanobeir_ne / NanoMSMARCO_cosine_recall@5	0.6600	0.5200	0.7879
nanobeir_ne / NanoMSMARCO_cosine_recall@10	0.7400	0.6800	0.9189
nanobeir_ne / NanoMSMARCO_cosine_ndcg@10	0.4955	0.4125	0.8325
nanobeir_ne / NanoMSMARCO_cosine_mrr@10	0.4174	0.3298	0.7901
nanobeir_ne / NanoMSMARCO_cosine_map@100	0.4252	0.3412	0.8024
nanobeir_ne / NanoNFCorpus_cosine_accuracy@1	0.2800	0.1600	0.5714
nanobeir_ne / NanoNFCorpus_cosine_accuracy@3	0.4400	0.3400	0.7727
nanobeir_ne / NanoNFCorpus_cosine_accuracy@5	0.4400	0.4800	1.0909
nanobeir_ne / NanoNFCorpus_cosine_accuracy@10	0.4400	0.5600	1.2727
nanobeir_ne / NanoNFCorpus_cosine_precision@1	0.2800	0.1600	0.5714
nanobeir_ne / NanoNFCorpus_cosine_precision@3	0.2600	0.1933	0.7436
nanobeir_ne / NanoNFCorpus_cosine_precision@5	0.2120	0.2240	1.0566
nanobeir_ne / NanoNFCorpus_cosine_precision@10	0.1600	0.1840	1.1500
nanobeir_ne / NanoNFCorpus_cosine_recall@1	0.0084	0.0054	0.6468
nanobeir_ne / NanoNFCorpus_cosine_recall@3	0.0413	0.0296	0.7165
nanobeir_ne / NanoNFCorpus_cosine_recall@5	0.0486	0.0483	0.9939
nanobeir_ne / NanoNFCorpus_cosine_recall@10	0.0617	0.0802	1.3003
nanobeir_ne / NanoNFCorpus_cosine_ndcg@10	0.1975	0.1976	1.0005
nanobeir_ne / NanoNFCorpus_cosine_mrr@10	0.3500	0.2882	0.8235
nanobeir_ne / NanoNFCorpus_cosine_map@100	0.0701	0.0660	0.9418
nanobeir_ne / NanoNQ_cosine_accuracy@1	0.2000	0.1600	0.8000
nanobeir_ne / NanoNQ_cosine_accuracy@3	0.3400	0.3000	0.8824
nanobeir_ne / NanoNQ_cosine_accuracy@5	0.3400	0.3200	0.9412
nanobeir_ne / NanoNQ_cosine_accuracy@10	0.4400	0.4600	1.0455
nanobeir_ne / NanoNQ_cosine_precision@1	0.2000	0.1600	0.8000
nanobeir_ne / NanoNQ_cosine_precision@3	0.1133	0.1000	0.8824
nanobeir_ne / NanoNQ_cosine_precision@5	0.0680	0.0640	0.9412
nanobeir_ne / NanoNQ_cosine_precision@10	0.0440	0.0460	1.0455
nanobeir_ne / NanoNQ_cosine_recall@1	0.1800	0.1500	0.8333
nanobeir_ne / NanoNQ_cosine_recall@3	0.3100	0.2800	0.9032
nanobeir_ne / NanoNQ_cosine_recall@5	0.3100	0.3000	0.9677
nanobeir_ne / NanoNQ_cosine_recall@10	0.4100	0.4200	1.0244
nanobeir_ne / NanoNQ_cosine_ndcg@10	0.2951	0.2817	0.9545
nanobeir_ne / NanoNQ_cosine_mrr@10	0.2767	0.2495	0.9016
nanobeir_ne / NanoNQ_cosine_map@100	0.2685	0.2439	0.9087
nanobeir_ne / NanoQuoraRetrieval_cosine_accuracy@1	0.8200	0.7400	0.9024
nanobeir_ne / NanoQuoraRetrieval_cosine_accuracy@3	0.9000	0.8200	0.9111
nanobeir_ne / NanoQuoraRetrieval_cosine_accuracy@5	0.9200	0.8600	0.9348
nanobeir_ne / NanoQuoraRetrieval_cosine_accuracy@10	0.9800	0.9000	0.9184
nanobeir_ne / NanoQuoraRetrieval_cosine_precision@1	0.8200	0.7400	0.9024
nanobeir_ne / NanoQuoraRetrieval_cosine_precision@3	0.3533	0.3000	0.8491
nanobeir_ne / NanoQuoraRetrieval_cosine_precision@5	0.2360	0.2040	0.8644
nanobeir_ne / NanoQuoraRetrieval_cosine_precision@10	0.1320	0.1140	0.8636
nanobeir_ne / NanoQuoraRetrieval_cosine_recall@1	0.7240	0.6773	0.9355
nanobeir_ne / NanoQuoraRetrieval_cosine_recall@3	0.8380	0.7713	0.9204
nanobeir_ne / NanoQuoraRetrieval_cosine_recall@5	0.8860	0.8260	0.9323
nanobeir_ne / NanoQuoraRetrieval_cosine_recall@10	0.9660	0.8727	0.9034
nanobeir_ne / NanoQuoraRetrieval_cosine_ndcg@10	0.8797	0.7971	0.9061
nanobeir_ne / NanoQuoraRetrieval_cosine_mrr@10	0.8707	0.7867	0.9035
nanobeir_ne / NanoQuoraRetrieval_cosine_map@100	0.8452	0.7688	0.9096
nanobeir_ne / NanoSCIDOCS_cosine_accuracy@1	0.2000	0.1600	0.8000
nanobeir_ne / NanoSCIDOCS_cosine_accuracy@3	0.3600	0.2800	0.7778
nanobeir_ne / NanoSCIDOCS_cosine_accuracy@5	0.4600	0.4200	0.9130
nanobeir_ne / NanoSCIDOCS_cosine_accuracy@10	0.5800	0.5000	0.8621
nanobeir_ne / NanoSCIDOCS_cosine_precision@1	0.2000	0.1600	0.8000
nanobeir_ne / NanoSCIDOCS_cosine_precision@3	0.1733	0.1267	0.7308
nanobeir_ne / NanoSCIDOCS_cosine_precision@5	0.1480	0.1200	0.8108
nanobeir_ne / NanoSCIDOCS_cosine_precision@10	0.0980	0.0840	0.8571
nanobeir_ne / NanoSCIDOCS_cosine_recall@1	0.0420	0.0330	0.7857
nanobeir_ne / NanoSCIDOCS_cosine_recall@3	0.1080	0.0770	0.7130
nanobeir_ne / NanoSCIDOCS_cosine_recall@5	0.1557	0.1240	0.7966
nanobeir_ne / NanoSCIDOCS_cosine_recall@10	0.2047	0.1730	0.8453
nanobeir_ne / NanoSCIDOCS_cosine_ndcg@10	0.1883	0.1565	0.8312
nanobeir_ne / NanoSCIDOCS_cosine_mrr@10	0.3002	0.2527	0.8416
nanobeir_ne / NanoSCIDOCS_cosine_map@100	0.1343	0.1044	0.7773
nanobeir_ne / NanoArguAna_cosine_accuracy@1	0.1200	0.0800	0.6667
nanobeir_ne / NanoArguAna_cosine_accuracy@3	0.5200	0.4000	0.7692
nanobeir_ne / NanoArguAna_cosine_accuracy@5	0.5800	0.5200	0.8966
nanobeir_ne / NanoArguAna_cosine_accuracy@10	0.7400	0.6200	0.8378
nanobeir_ne / NanoArguAna_cosine_precision@1	0.1200	0.0800	0.6667
nanobeir_ne / NanoArguAna_cosine_precision@3	0.1733	0.1333	0.7692
nanobeir_ne / NanoArguAna_cosine_precision@5	0.1160	0.1040	0.8966
nanobeir_ne / NanoArguAna_cosine_precision@10	0.0740	0.0620	0.8378
nanobeir_ne / NanoArguAna_cosine_recall@1	0.1200	0.0800	0.6667
nanobeir_ne / NanoArguAna_cosine_recall@3	0.5200	0.4000	0.7692
nanobeir_ne / NanoArguAna_cosine_recall@5	0.5800	0.5200	0.8966
nanobeir_ne / NanoArguAna_cosine_recall@10	0.7400	0.6200	0.8378
nanobeir_ne / NanoArguAna_cosine_ndcg@10	0.4276	0.3476	0.8130
nanobeir_ne / NanoArguAna_cosine_mrr@10	0.3276	0.2602	0.7945
nanobeir_ne / NanoArguAna_cosine_map@100	0.3367	0.2650	0.7870
nanobeir_ne / NanoSciFact_cosine_accuracy@1	0.3200	0.2400	0.7500
nanobeir_ne / NanoSciFact_cosine_accuracy@3	0.4800	0.4600	0.9583
nanobeir_ne / NanoSciFact_cosine_accuracy@5	0.5400	0.5200	0.9630
nanobeir_ne / NanoSciFact_cosine_accuracy@10	0.6600	0.5800	0.8788
nanobeir_ne / NanoSciFact_cosine_precision@1	0.3200	0.2400	0.7500
nanobeir_ne / NanoSciFact_cosine_precision@3	0.1667	0.1600	0.9600
nanobeir_ne / NanoSciFact_cosine_precision@5	0.1120	0.1080	0.9643
nanobeir_ne / NanoSciFact_cosine_precision@10	0.0700	0.0620	0.8857
nanobeir_ne / NanoSciFact_cosine_recall@1	0.3050	0.2250	0.7377
nanobeir_ne / NanoSciFact_cosine_recall@3	0.4600	0.4400	0.9565
nanobeir_ne / NanoSciFact_cosine_recall@5	0.5100	0.5000	0.9804
nanobeir_ne / NanoSciFact_cosine_recall@10	0.6250	0.5650	0.9040
nanobeir_ne / NanoSciFact_cosine_ndcg@10	0.4596	0.4007	0.8720
nanobeir_ne / NanoSciFact_cosine_mrr@10	0.4155	0.3574	0.8600
nanobeir_ne / NanoSciFact_cosine_map@100	0.4093	0.3533	0.8632
nanobeir_ne / NanoTouche2020_cosine_accuracy@1	0.3469	0.1224	0.3529
nanobeir_ne / NanoTouche2020_cosine_accuracy@3	0.5510	0.3469	0.6296
nanobeir_ne / NanoTouche2020_cosine_accuracy@5	0.6735	0.5510	0.8182
nanobeir_ne / NanoTouche2020_cosine_accuracy@10	0.8571	0.7143	0.8333
nanobeir_ne / NanoTouche2020_cosine_precision@1	0.3469	0.1224	0.3529
nanobeir_ne / NanoTouche2020_cosine_precision@3	0.3333	0.1633	0.4898
nanobeir_ne / NanoTouche2020_cosine_precision@5	0.3224	0.2122	0.6582
nanobeir_ne / NanoTouche2020_cosine_precision@10	0.2939	0.1857	0.6319
nanobeir_ne / NanoTouche2020_cosine_recall@1	0.0216	0.0091	0.4228
nanobeir_ne / NanoTouche2020_cosine_recall@3	0.0678	0.0325	0.4793
nanobeir_ne / NanoTouche2020_cosine_recall@5	0.1083	0.0700	0.6466
nanobeir_ne / NanoTouche2020_cosine_recall@10	0.1892	0.1182	0.6244
nanobeir_ne / NanoTouche2020_cosine_ndcg@10	0.3143	0.1835	0.5838
nanobeir_ne / NanoTouche2020_cosine_mrr@10	0.4881	0.2811	0.5759
nanobeir_ne / NanoTouche2020_cosine_map@100	0.2267	0.1326	0.5847
nanobeir_ne / NanoBEIR_mean_cosine_accuracy@1	0.3405	0.2479	0.7279
nanobeir_ne / NanoBEIR_mean_cosine_accuracy@3	0.5439	0.4313	0.7929
nanobeir_ne / NanoBEIR_mean_cosine_accuracy@5	0.6010	0.5224	0.8691
nanobeir_ne / NanoBEIR_mean_cosine_accuracy@10	0.6952	0.6226	0.8957
nanobeir_ne / NanoBEIR_mean_cosine_precision@1	0.3405	0.2479	0.7279
nanobeir_ne / NanoBEIR_mean_cosine_precision@3	0.2318	0.1751	0.7555
nanobeir_ne / NanoBEIR_mean_cosine_precision@5	0.1790	0.1474	0.8237
nanobeir_ne / NanoBEIR_mean_cosine_precision@10	0.1263	0.1046	0.8281
nanobeir_ne / NanoBEIR_mean_cosine_recall@1	0.2008	0.1571	0.7826
nanobeir_ne / NanoBEIR_mean_cosine_recall@3	0.3514	0.2910	0.8281
nanobeir_ne / NanoBEIR_mean_cosine_recall@5	0.4005	0.3432	0.8570
nanobeir_ne / NanoBEIR_mean_cosine_recall@10	0.4847	0.4161	0.8585
nanobeir_ne / NanoBEIR_mean_cosine_ndcg@10	0.4122	0.3388	0.8218
nanobeir_ne / NanoBEIR_mean_cosine_mrr@10	0.4542	0.3612	0.7953
nanobeir_ne / NanoBEIR_mean_cosine_map@100	0.3419	0.2769	0.8098

Citation

If you use this model or the pruning approach, please cite:

@misc{subedi2025tokenpruning,
  author = {Sanjaya Subedi},
  title  = {Token Embedding Pruning for Sentence Transformers},
  year   = {2026},
  note   = {Available at: https://sanjayasubedi.com.np/deeplearning/shrinking-embedding-models-by-pruning-vocabulary/}
}