e5-base Open-Source Text Embedding Model - Free for Natural Language Processing Tasks such as Classification and Retrieval

E5 Base

Developed by intfloat

E5-base is a general-purpose text embedding model suitable for various natural language processing tasks such as classification, retrieval, clustering, and semantic similarity calculation.

Text Embedding

Safetensors

EnglishOpen Source License:MIT #Multitask Text Embedding #High-Precision Classification #Semantic Retrieval Optimization

Downloads 30.85k

Release Time : 12/26/2022

Model Overview

E5-base is a Transformer-based text embedding model capable of converting text into high-dimensional vector representations, suitable for various downstream tasks.

Model Features

Multitask Support

Supports various natural language processing tasks, including classification, retrieval, clustering, and semantic similarity calculation.

High Performance

Performs excellently on multiple benchmark datasets, such as the MTEB dataset.

Versatility

Applicable to various text processing scenarios without requiring extensive task-specific adjustments.

Model Capabilities

Text classification

Text retrieval

Text clustering

Semantic similarity calculation

Text reranking

Use Cases

E-commerce

Product Review Classification

Classify Amazon product reviews to identify positive and negative feedback.

Achieves an accuracy of 87.96% on the MTEB AmazonPolarityClassification dataset.

Product Retrieval

Retrieve relevant products based on user queries.

Achieves an F1 score of 42.23 on the MTEB AmazonReviewsClassification dataset.

Academic Research

Paper Clustering

Cluster academic papers from arXiv and BioRxiv.

Achieves a V-measure of 44.57 on the MTEB ArxivClusteringP2P dataset.

Q&A Systems

Duplicate Question Detection

Detect duplicate questions in Q&A communities.

Achieves a MAP of 59.66 on the MTEB AskUbuntuDupQuestions dataset.

🚀 e5-base

This is the e5-base model, which has been tested on multiple tasks in the MTEB benchmark, including Classification, Retrieval, Clustering, Reranking, and STS. The following shows detailed performance metrics on various datasets.

📚 Documentation

Model Information

Property	Details
Model Name	e5-base
Tags	mteb, Sentence Transformers, sentence-similarity, sentence-transformers

Performance Metrics

Classification Tasks

Dataset	Accuracy	AP	F1
MTEB AmazonCounterfactualClassification (en)	79.71641791044777	44.15426065428253	73.89474407693241
MTEB AmazonPolarityClassification	87.9649	84.10171551915973	87.94148377827356
MTEB AmazonReviewsClassification (en)	42.645999999999994	-	42.230574673549
MTEB Banking77Classification	83.32792207792208	-	83.32037485050513

Retrieval Tasks

Dataset	MAP@1	MAP@10	MAP@100	MAP@1000	MRR@1	MRR@10	MRR@100	MRR@1000	NDCG@1	NDCG@10	NDCG@100	NDCG@1000	Precision@1	Precision@10	Precision@100	Precision@1000	Recall@1	Recall@10	Recall@100	Recall@1000
MTEB ArguAna	26.814	42.681999999999995	43.714	43.724000000000004	27.168999999999997	42.84	43.864	43.875	26.814	51.410999999999994	55.713	55.957	26.814	7.922999999999999	0.9780000000000001	0.1	26.814	79.232	97.795	99.644
MTEB CQADupstackAndroidRetrieval	30.659	40.333999999999996	41.763	41.894	37.482	45.736	46.591	46.644999999999996	37.482	45.606	51.172	53.407000000000004	37.482	8.254999999999999	1.3719999999999999	0.186	30.659	55.409	78.687	93.068
MTEB CQADupstackEnglishRetrieval	30.977	40.296	41.453	41.581	39.108	46.894000000000005	47.55	47.598	39.108	45.717	49.941	52.138	39.108	8.306	1.3419999999999999	0.184	30.977	54.688	72.556	86.53800000000001
MTEB CQADupstackGamingRetrieval	39.812	50.1	51.193999999999996	51.258	45.266	53.459999999999994	54.19199999999999	54.228	45.266	55.034000000000006	59.458	60.862	45.266	8.483	1.162	0.133	39.812	66.36	85.392	95.523
MTEB CQADupstackGisRetrieval	26.186	33.18	34.052	34.149	28.136	35.195	35.996	36.076	28.136	37.275999999999996	41.935	44.389	28.136	5.457999999999999	0.826	0.107	26.186	47.961999999999996	70.072	88.505
MTEB CQADupstackMathematicaRetrieval	16.586000000000002	23.637	24.82	24.95	20.771	27.839999999999996	28.887	28.967	20.771	28.255000000000003	33.886	36.963	20.771	5.1	0.9119999999999999	0.132	16.586000000000002	38.456	62.666	-

Clustering Tasks

Dataset	V-Measure
MTEB ArxivClusteringP2P	44.56933066536439
MTEB ArxivClusteringS2S	40.47647746165173
MTEB BiorxivClusteringP2P	36.18605446588703
MTEB BiorxivClusteringS2S	32.72379130181917

Reranking Tasks

Dataset	MAP	MRR
MTEB AskUbuntuDupQuestions	59.65675531567043	72.95255683067317

STS Tasks

Dataset	Cosine Similarity Pearson	Cosine Similarity Spearman	Euclidean Pearson	Euclidean Spearman	Manhattan Pearson	Manhattan Spearman
MTEB BIOSSES	85.83147014162338	85.1031439521441	83.53609085510973	84.59650590202833	83.14611947586386	84.13384475757064

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