GIST-all-MiniLM-L6-v2 Open-source Model - Handles sentence similarity and is suitable for various NLP scenarios

GIST All MiniLM L6 V2

Developed by avsolatorio

GIST-all-MiniLM-L6-v2 is a Sentence-Transformers model focused on sentence similarity tasks, suitable for various natural language processing scenarios such as feature extraction, classification, retrieval, and clustering.

Text Embedding EnglishOpen Source License:MIT #Sentence similarity calculation #Multi-task performance balancing #Text retrieval optimization

Downloads 99.17k

Release Time : 2/3/2024

Model Overview

This model is based on the MiniLM architecture and has been optimized to handle sentence-level similarity calculations, supporting multiple downstream tasks such as text classification, information retrieval, and clustering analysis.

Model Features

Efficient sentence embedding

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

Multi-task adaptation

Optimized to adapt to multiple downstream tasks, including classification, retrieval, and clustering.

Lightweight architecture

Based on the MiniLM architecture, reducing computational resource requirements while maintaining performance.

Model Capabilities

Sentence similarity calculation

Text feature extraction

Text classification

Information retrieval

Text clustering

Semantic text similarity analysis

Use Cases

E-commerce

Product review classification

Classify the sentiment polarity of Amazon product reviews

Achieved 87.19% accuracy on the AmazonPolarityClassification dataset

Counterfactual review detection

Identify counterfactual reviews on the Amazon platform

Achieved 72.89% accuracy on the AmazonCounterfactualClassification dataset

Finance

Bank customer service question classification

Automatically classify bank customer service questions

Achieved 84.24% accuracy on the Banking77Classification dataset

Academic research

Academic paper clustering

Perform topic clustering on arXiv and bioRxiv papers

Achieved a 45.31 V-measure on the ArxivClusteringP2P dataset

🚀 GIST-all-MiniLM-L6-v2

This is the GIST-all-MiniLM-L6-v2 model from the sentence-transformers library, which is mainly used for sentence similarity tasks. It has been tested on multiple datasets and shows performance across various tasks such as Classification, Retrieval, Clustering, Reranking, and STS.

📚 Documentation

Model Information

Property	Details
Model Type	GIST-all-MiniLM-L6-v2
Library Name	sentence-transformers
Pipeline Tag	sentence-similarity
Tags	feature-extraction, mteb, sentence-similarity, sentence-transformers
License	MIT

Performance Results

The model has been evaluated on multiple datasets for different tasks. Here are the detailed results:

Classification Tasks

Task	Dataset	Accuracy	AP	F1
Classification	MTEB AmazonCounterfactualClassification (en)	72.8955223880597	35.447605103320775	66.82951715365854
Classification	MTEB AmazonPolarityClassification	87.19474999999998	83.09577890808514	87.13833121762009
Classification	MTEB AmazonReviewsClassification (en)	42.556000000000004	-	42.236256693772276
Classification	MTEB Banking77Classification	84.2435064935065	-	84.2334859253828

Retrieval Tasks

Task	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
Retrieval	MTEB ArguAna	26.884999999999998	42.364000000000004	43.382	43.391000000000005	26.884999999999998	42.193999999999996	43.211	43.221	26.884999999999998	51.254999999999995	55.481	55.68300000000001	26.884999999999998	7.9799999999999995	0.98	0.1	26.884999999999998	79.801	98.009	99.502
Retrieval	MTEB CQADupstackAndroidRetrieval	35.016999999999996	47.019	48.634	48.757	43.491	53.284	54.038	54.071000000000005	43.491	53.498999999999995	58.733999999999995	60.307	43.491	10.315000000000001	1.6209999999999998	0.20500000000000002	35.016999999999996	64.92	86.605	96.174
Retrieval	MTEB CQADupstackEnglishRetrieval	29.866	40.438	41.77	41.913	37.834	46.765	47.410000000000004	47.461	37.834	46.303	50.879	53.112	37.834	8.898	1.4409999999999998	0.19499999999999998	29.866	56.06100000000001	75.809	89.875
Retrieval	MTEB CQADupstackGamingRetrieval	38.985	51.165000000000006	52.17	52.229000000000006	44.577	54.493	55.137	55.167	44.577	56.825	60.842	62.015	44.577	9.11	1.206	0.135	38.985	70.164	87.708	95.979
Retrieval	MTEB CQADupstackGisRetrieval	28.137	36.729	37.851	37.932	30.621	39.007	39.961	40.02	30.621	41.772	47.181	49.053999999999995	30.621	6.372999999999999	0.955	0.11499999999999999	28.137	55.162	79.931	93.67
Retrieval	MTEB CQADupstackMathematicaRetrieval	16.798	25.267	26.579000000000004	26.697	20.771	29.843999999999998	30.849	30.916	20.771	30.792	36.945	39.619	20.771	5.734	1.031	0.13899999999999998	16.798	43.332	70.016	88.90400000000001
Retrieval	MTEB CQADupstackPhysicsRetrieval	31.180000000000003	41.78	43.102000000000004	43.222	37.824999999999996	47.481	48.268	48.313	37.824999999999996	47.827	53.407000000000004	55.321	37.824999999999996	8.652999999999999	1.354	0.172	31.180000000000003	59.894000000000005	83.722	95.705
Retrieval	MTEB CQADupstackProgrammersRetrieval	24.66	34.141	35.478	35.594	29.909000000000002	38.949	39.803	39.867999999999995	29.909000000000002	40.012	45.707	48.15	29.909000000000002	7.693999999999999	1.2229999999999999	0.16	24.66	52.478	77.051	93.872
Retrieval	MTEB CQADupstackRetrieval	26.768416666666667	36.2485	37.520833333333336	37.64033333333334	31.65408333333334	40.43866666666667	41.301249999999996	41.357499999999995	31.65408333333334	41.76983333333334	47.138	49.33816666666667	31.65408333333334	7.396249999999998	1.1974166666666666	0.15791666666666668	26.768416666666667	53.82366666666667	77.39600000000002	92.46300000000001
Retrieval	MTEB CQADupstackStatsRetrieval	24.369	32.025	33.08	33.169	27.301	34.64	35.556	35.616	27.301	36.386	41.598	43.864999999999995	27.301	5.782	0.923	0.11900000000000001	24.369	47.026	70.76400000000001	87.705

Clustering Tasks

Task	Dataset	V-Measure
Clustering	MTEB ArxivClusteringP2P	45.31044837358167
Clustering	MTEB ArxivClusteringS2S	35.44751738734691
Clustering	MTEB BiorxivClusteringP2P	38.38358435972693
Clustering	MTEB BiorxivClusteringS2S	31.093619653843124

Reranking Task

Task	Dataset	MAP	MRR
Reranking	MTEB AskUbuntuDupQuestions	62.96517580629869	76.30051004704744

STS Task

Task	Dataset	Cos Sim Pearson	Cos Sim Spearman	Euclidean Pearson	Euclidean Spearman	Manhattan Pearson	Manhattan Spearman
STS	MTEB BIOSSES	83.97262600499639	81.25787561220484	64.96260261677082	64.17616109254686	65.05620628102835	64.71171546419122

📄 License

This project is licensed under the MIT License.

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