BGE-micro-v2 Open-Source Sentence Embedding Model - Freely Perform Sentence Similarity Calculation and Feature Extraction

Bge Micro V2

Developed by SmartComponents

bge_micro is a sentence embedding model based on sentence-transformers, focusing on sentence similarity calculation and feature extraction tasks.

Text Embedding

Transformers

Downloads 468

Release Time : 2/15/2024

Model Overview

This model is mainly used to generate sentence embeddings for tasks such as sentence similarity calculation, text classification, and information retrieval.

Model Features

Efficient sentence embedding

Capable of quickly generating high-quality sentence embeddings, suitable for large-scale text processing.

Multi-task support

Supports various natural language processing tasks such as sentence similarity calculation, feature extraction, and text classification.

Good performance on MTEB benchmarks

Achieved commendable results in multiple MTEB benchmark tasks, demonstrating its strong generalization capability.

Model Capabilities

Sentence similarity calculation

Feature extraction

Text classification

Information retrieval

Clustering analysis

Use Cases

Text classification

Amazon review classification

Used for sentiment or topic classification of Amazon product reviews.

Achieved 79.75% accuracy in the MTEB AmazonPolarityClassification task.

Bank customer service question classification

Used for automatic classification of user questions in bank customer service systems.

Achieved 81.17% accuracy in the MTEB Banking77Classification task.

Information retrieval

Q&A system retrieval

Used to retrieve the most relevant answers from a knowledge base based on user questions.

Achieved map@100 of 39.47 in the MTEB CQADupstackAndroidRetrieval task.

Sentence similarity

Duplicate question detection

Used to identify duplicate questions in forums or Q&A platforms.

Achieved mrr of 71.94 in the MTEB AskUbuntuDupQuestions task.

Biomedical text similarity

Used to calculate semantic similarity between biomedical texts.

Achieved cos_sim_pearson of 84.16 in the MTEB BIOSSES task.

🚀 BGE Micro

This is a model for sentence similarity tasks. It belongs to the category of sentence-transformers and can be used for feature extraction and various natural language processing tasks.

📚 Documentation

Model Information

Property	Details
Pipeline Tag	sentence-similarity
Tags	sentence-transformers, feature-extraction, sentence-similarity, transformers, mteb
Model Name	bge_micro

Results on Multiple Datasets

1. Classification Tasks

Dataset	Accuracy	AP	F1
MTEB AmazonCounterfactualClassification (en)	67.76119402985074	29.637849284211114	61.31181187111905
MTEB AmazonPolarityClassification	79.7547	74.21401629809145	79.65319615433783
MTEB AmazonReviewsClassification (en)	37.452000000000005	-	37.0245198854966
MTEB Banking77Classification	81.17207792207792	-	81.09893836310513

2. 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	31.152	46.702	47.563	47.567	32.006	47.064	47.910000000000004	47.915	31.152	55.308	58.965	59.067	31.152	8.279	0.987	0.1	31.152	82.788	98.72	99.502
MTEB CQADupstackAndroidRetrieval	28.233999999999998	38.092999999999996	39.473	39.614	35.193000000000005	44.089	44.927	44.988	35.193000000000005	44.04	49.262	51.847	35.193000000000005	8.555	1.3820000000000001	0.189	28.233999999999998	55.094	76.85300000000001	94.163
MTEB CQADupstackEnglishRetrieval	21.538	28.449	29.471000000000004	29.599999999999998	26.815	33.331	34.114	34.182	26.815	32.67	37.039	39.769	26.815	5.955	1.02	0.152	21.538	40.186	58.948	77.158
MTEB CQADupstackGamingRetrieval	35.211999999999996	46.562	47.579	47.646	40.627	49.928	50.647	50.685	40.627	52.217	56.423	57.821999999999996	40.627	8.476	1.15	0.132	35.211999999999996	65.692	84.011	94.03099999999999
MTEB CQADupstackGisRetrieval	22.09	29.516	30.462	30.56	23.616	31.221	32.057	32.137	23.616	33.97	38.806000000000004	41.393	23.616	5.299	0.812	0.107	22.09	46.089999999999996	68.729	88.435
MTEB CQADupstackMathematicaRetrieval	15.469	22.436	23.465	23.608999999999998	18.905	26.55	27.46	27.553	18.905	27.437	32.555	35.885	18.905	5.2490000000000006	0.889	0.131	15.469	38.706	61.242	84.84

3. Clustering Tasks

Dataset	V-Measure
MTEB ArxivClusteringP2P	44.52692241938116
MTEB ArxivClusteringS2S	33.245710292773595
MTEB BiorxivClusteringP2P	36.109308463095516
MTEB BiorxivClusteringS2S	28.06048212317168

4. Reranking Task

Dataset	MAP	MRR
MTEB AskUbuntuDupQuestions	58.08493637155168	71.94378490084861

5. STS Task

Dataset	Cosine Similarity Pearson	Cosine Similarity Spearman	Euclidean Pearson	Euclidean Spearman	Manhattan Pearson	Manhattan Spearman
MTEB BIOSSES	84.1602804378326	82.92478106365587	82.27930167277077	82.18560759458093	82.34277425888187	81.72776583704467

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