intfloat/multilingual-e5-large

Multilingual E5 Text Embeddings: A Technical Report. Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei, arXiv 2024

Overview

Architecture

XLM-RoBERTa

Parameters

560M

Tasks

Encode

Outputs

Dense

Dimensions

Dense: 1,024

Max Sequence Length

512 tokens

License

mit

Languages

multilingual, af, am, ar, as, az, be, bg, bn, br, bs, ca, cs, cy, da, de, el, en, eo, es, et, eu, fa, fi, fr, fy, ga, gd, gl, gu, ha, he, hi, hr, hu, hy, id, is, it, ja, jv, ka, kk, km, kn, ko, ku, ky, la, lo, lt, lv, mg, mk, ml, mn, mr, ms, my, ne, nl, no, om, or, pa, pl, ps, pt, ro, ru, sa, sd, si, sk, sl, so, sq, sr, su, sv, sw, ta, te, th, tl, tr, ug, uk, ur, uz, vi, xh, yi, zh

View on HuggingFace →

Benchmarks

CQADupstackPhysicsRetrieval

scientific retrieval en

Duplicate question retrieval from StackExchange Physics

Corpus: 38,314 Queries: 1,039

default

Quality

map at 10 0.3783

mrr at 10 0.4371

ndcg at 10 0.4367

Performance L4 b1 c16

Corpus 23.7K tok/s

Corpus p50 85.2ms

Query 2.9K tok/s

Query p50 58.0ms

sentence_transformer

Quality

map at 10 0.3178

mrr at 10 0.3695

ndcg at 10 0.3724

Reference →

CosQA

technology retrieval en

Code search with natural language queries

Corpus: 6,267 Queries: 500

default

Quality

map at 10 0.2575

mrr at 10 0.2840

ndcg at 10 0.3378

Performance L4 b1 c16

Corpus 14.2K tok/s

Corpus p50 60.8ms

Query 1.8K tok/s

Query p50 54.9ms

sentence_transformer

Quality

map at 10 0.2630

mrr at 10 0.2886

ndcg at 10 0.3349

Reference →

FiQA2018

finance retrieval en

Financial opinion mining and question answering

Corpus: 57,599 Queries: 648

Quality

ndcg at 10 0.3698

map at 10 0.2966

mrr at 10 0.4506

Performance L4 b1 c16

Corpus 29.8K tok/s

Corpus p50 89.9ms

Query 2.8K tok/s

Query p50 63.0ms

Reference →

LegalBenchConsumerContractsQA

legal retrieval en

Question answering on consumer contracts

Corpus: 153 Queries: 396

default

Quality

map at 10 0.6751

mrr at 10 0.6743

ndcg at 10 0.7332

Performance L4 b1 c16

Corpus 46.3K tok/s

Corpus p50 173.9ms

Query 3.9K tok/s

Query p50 62.7ms

sentence_transformer

Quality

map at 10 0.6685

mrr at 10 0.6678

ndcg at 10 0.7251

Reference →

NFCorpus

medical retrieval en

Biomedical literature search from NutritionFacts.org

Corpus: 3,593 Queries: 323

Quality

ndcg at 10 0.3063

map at 10 0.1089

mrr at 10 0.4947

Performance L4 b1 c16

Corpus 33.8K tok/s

Corpus p50 142.5ms

Query 1.4K tok/s

Query p50 55.6ms

Reference →

NanoFiQA2018Retrieval

finance retrieval en

Smaller subset of the FiQA financial QA dataset

Quality

ndcg at 10 0.5035

map at 10 0.4111

mrr at 10 0.5364

Performance L4 b1 c16

Corpus 27.7K tok/s

Corpus p50 91.9ms

Query 3.1K tok/s

Query p50 55.7ms

Reference →

SCIDOCS

scientific retrieval en

Citation prediction, document classification, and recommendation for scientific papers

Corpus: 25,656 Queries: 1,000

Quality

ndcg at 10 0.1491

map at 10 0.0844

mrr at 10 0.2717

Performance L4 b1 c16

Corpus 27.5K tok/s

Corpus p50 108.6ms

Query 3.1K tok/s

Query p50 56.9ms

Reference →

SciFact

scientific retrieval en

Scientific claim verification using research literature

Corpus: 5,183 Queries: 300

Quality

ndcg at 10 0.6926

map at 10 0.6515

mrr at 10 0.6659

Performance L4 b1 c16

Corpus 33.6K tok/s

Corpus p50 134.7ms

Query 4.3K tok/s

Query p50 60.0ms

Reference →

StackOverflowQA

technology retrieval en

Programming question answering from Stack Overflow

Corpus: 19,931 Queries: 1,994

default

Quality

map at 10 0.8719

mrr at 10 0.8719

ndcg at 10 0.8889

Performance L4 b1 c16

Corpus 33.7K tok/s

Corpus p50 112.0ms

Query 38.8K tok/s

Query p50 121.7ms

sentence_transformer

Quality

map at 10 0.8606

mrr at 10 0.8606

ndcg at 10 0.8774

Reference →

Overview

Benchmarks

CQADupstackPhysicsRetrieval

CosQA

FiQA2018

LegalBenchConsumerContractsQA

NFCorpus

NanoFiQA2018Retrieval

SCIDOCS

SciFact

StackOverflowQA

Self-hosted inference for search & document processing