Multilingual E5 Large Open-source Model - Supports Multilingual Sentence Similarity Comparison and Feature Extraction

Multilingual E5 Large Pooled

Developed by Hiveurban

Multilingual E5 Large is a multilingual sentence transformer model focused on sentence similarity and feature extraction tasks, supporting multiple languages.

Text Embedding

Safetensors

Supports Multiple LanguagesOpen Source License:MIT #Multilingual Text Embedding #Cross-Language Retrieval #High-Precision Classification

Downloads 3,803

Release Time : 9/24/2024

Model Overview

This model is a multilingual model based on sentence-transformers, primarily used for sentence similarity calculation and feature extraction, performing well in classification, retrieval, and clustering tasks across multiple languages.

Model Features

Multilingual Support

Supports processing in multiple languages, including English, German, Spanish, French, Japanese, and Chinese.

Multifunctional Task Handling

Capable of handling various natural language processing tasks, including classification, retrieval, clustering, and sentence similarity calculation.

High Performance

Performs excellently on multiple standard datasets, achieving 93.49% accuracy in the Amazon Polarity Classification task.

Model Capabilities

Sentence similarity calculation

Text feature extraction

Multilingual text classification

Information retrieval

Text clustering

Bilingual text mining

Use Cases

E-commerce

Product Review Classification

Performs sentiment polarity classification on Amazon product reviews

Achieved 93.49% accuracy in the Amazon Polarity Classification task

Multilingual Review Analysis

Processes user reviews in multiple languages

Performs well in classification tasks for English, German, Spanish, French, Japanese, and Chinese reviews

Information Retrieval

Document Retrieval

Used to retrieve relevant documents or paragraphs

Performs excellently in the ArguAna retrieval task

Academic Research

Paper Clustering

Performs topic clustering on academic papers

Achieved a v_measure of 44.31% in the ArxivClustering task

🚀 multilingual-e5-large

This is a model related to sentence processing, which can be used for tasks such as sentence similarity calculation, feature extraction, etc. It has been tested on multiple datasets and shown certain performance.

📚 Documentation

Model Information

Property	Details
Tags	mteb, Sentence Transformers, sentence-similarity, feature-extraction, sentence-transformers
Model Name	multilingual-e5-large

Performance Metrics

Classification Tasks

Dataset	Language	Accuracy	AP	F1
MTEB AmazonCounterfactualClassification	en	79.05970149253731	43.486574390835635	73.32700092140148
MTEB AmazonCounterfactualClassification	de	71.22055674518201	81.55756710830498	69.28271787752661
MTEB AmazonCounterfactualClassification	en-ext	80.41979010494754	29.34879922376344	67.62475449011278
MTEB AmazonCounterfactualClassification	ja	77.8372591006424	26.557560591210738	64.96619417368707
MTEB AmazonPolarityClassification	default	93.489875	90.98758636917603	93.48554819717332
MTEB AmazonReviewsClassification	en	47.564	-	46.75122173518047
MTEB AmazonReviewsClassification	de	45.400000000000006	-	44.17195682400632
MTEB AmazonReviewsClassification	es	43.068	-	42.38155696855596
MTEB AmazonReviewsClassification	fr	41.89	-	40.84407321682663
MTEB AmazonReviewsClassification	ja	40.120000000000005	-	39.522976223819114
MTEB AmazonReviewsClassification	zh	38.832	-	38.0392533394713
MTEB Banking77Classification	default	84.72727272727273	-	84.67672206031433

Retrieval Tasks

Dataset	MAP@1	MAP@10	MAP@100	MAP@1000	MAP@3	MAP@5	MRR@1	MRR@10	MRR@100	MRR@1000	MRR@3	MRR@5	NDCG@1	NDCG@10	NDCG@100	NDCG@1000	NDCG@3	NDCG@5	Precision@1	Precision@10	Precision@100	Precision@1000	Precision@3	Precision@5	Recall@1	Recall@10	Recall@100	Recall@1000	Recall@3	Recall@5
MTEB ArguAna	30.725	46.055	46.900999999999996	46.911	41.548	44.297	31.152	46.231	47.07	47.08	41.738	44.468999999999994	30.725	54.379999999999995	58.138	58.389	45.156	50.123	30.725	8.087	0.9769999999999999	0.1	18.54	13.542000000000002	30.725	80.868	97.653	99.57300000000001	55.619	67.71000000000001
MTEB CQADupstackRetrieval	25.558249999999997	34.44425000000001	35.59833333333333	35.706916666666665	31.691749999999995	33.252916666666664	30.252666666666666	38.60675	39.42666666666666	39.48408333333334	36.17441666666665	37.56275	30.252666666666666	39.683	44.68541666666667	46.94316666666668	34.961749999999995	37.215666666666664	30.252666666666666	6.904166666666667	1.0989999999999995	0.14733333333333334	16.037666666666667	11.413583333333333	25.558249999999997	51.13341666666666	73.08366666666667	88.79483333333334	37.989083333333326	43.787833333333325
MTEB ClimateFEVER	10.338	18.360000000000003	19.942	20.134	15.174000000000001	16.830000000000002	23.257	33.768	34.707	34.766000000000005	30.977	32.528	23.257	25.733	32.288	35.992000000000004	20.866	22.612	23.257	8.124	1.518	0.219	15.679000000000002	12.117	10.338	31.154	54.16	-	-	-

Clustering Tasks

Dataset	V-Measure
MTEB ArxivClusteringP2P	44.30960650674069
MTEB ArxivClusteringS2S	38.427074197498996
MTEB BiorxivClusteringP2P	35.34220182511161
MTEB BiorxivClusteringS2S	33.4987096128766

Reranking Tasks

Dataset	MAP	MRR
MTEB AskUbuntuDupQuestions	60.28270056031872	74.38332673789738

STS Tasks

Dataset	Cosine Similarity Pearson	Cosine Similarity Spearman	Euclidean Pearson	Euclidean Spearman	Manhattan Pearson	Manhattan Spearman
MTEB BIOSSES	84.05942144105269	82.51212105850809	81.95639829909122	82.3717564144213	81.79273425468256	82.20066817871039

Bitext Mining Tasks

Dataset	Language Pair	Accuracy	F1	Precision	Recall
MTEB BUCC	de-en	99.46764091858039	99.37717466945023	99.33194154488518	99.46764091858039
MTEB BUCC	fr-en	98.29407880255337	98.11248073959938	98.02443319392472	98.29407880255337
MTEB BUCC	ru-en	97.79009352268791	97.5176076665512	97.38136473848286	97.79009352268791
MTEB BUCC	zh-en	99.26276987888363	99.20133403545726	99.17500438827453	99.26276987888363

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご