GIST-small-Embedding-v0 Open-source Text Embedding Model - Optimize Retrieval Query Encoding Capability, Free to Use

GIST Small Embedding V0

Developed by avsolatorio

A text embedding model fine-tuned based on BAAI/bge-small-en-v1.5, trained with the MEDI dataset and MTEB classification task datasets, optimized for query encoding in retrieval tasks.

Text Embedding

Safetensors

EnglishOpen Source License:MIT #Instructionless Embedding #Cross-task Fine-tuning #Semantic Similarity Calculation

Downloads 945.68k

Release Time : 2/3/2024

Model Overview

This model generates embedding vectors without requiring instruction input, directly encoding query statements, suitable for text retrieval and similarity calculation tasks.

Model Features

No Instruction Input Required

Generates embedding vectors without constructing prompt statements, directly encoding queries.

Multi-dataset Training Integration

Fine-tuned with the MEDI dataset and MTEB classification task datasets to enhance model performance.

Optimized for Retrieval Tasks

Optimized for retrieval tasks, significantly improving performance in certain tasks.

Model Capabilities

Text Embedding Generation

Text Similarity Calculation

Retrieval Task Optimization

Use Cases

Information Retrieval

Document Retrieval

Used for quickly retrieving relevant documents or paragraphs.

Significant improvement observed in some MTEB tasks

Similarity Calculation

Text Similarity Analysis

Calculates the semantic similarity between two pieces of text.

🚀 GIST-small-Embedding-v0

GIST-small-Embedding-v0 is a model in the sentence-transformers library, which is mainly used for sentence similarity tasks. It has been tested on multiple datasets and shows good performance in various tasks such as classification, retrieval, clustering, and reranking.

📚 Documentation

Model Information

Property	Details
Library Name	sentence-transformers
Pipeline Tag	sentence-similarity
Tags	feature-extraction, mteb, sentence-similarity, sentence-transformers
Model Name	GIST-small-Embedding-v0
License	MIT

Performance Results

The following are the performance results of the GIST-small-Embedding-v0 model on different tasks and datasets:

Classification Tasks

Task	Dataset	Accuracy	AP	F1
Classification	MTEB AmazonCounterfactualClassification (en)	75.26865671641791	38.25623793370476	69.26434651320257
Classification	MTEB AmazonPolarityClassification	93.232225	89.97936072879344	93.22122653806187
Classification	MTEB AmazonReviewsClassification (en)	49.715999999999994	-	49.169789920136076
Classification	MTEB Banking77Classification	86.64935064935065	-	86.61903824934998

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	34.922	50.524	51.247	51.249	34.922	50.382000000000005	51.104000000000006	51.105999999999995	34.922	59.12	62.083999999999996	62.137	34.922	8.649	0.991	0.1	34.922	86.48599999999999	99.14699999999999	99.57300000000001
Retrieval	MTEB CQADupstackAndroidRetrieval	31.843	43.379	44.946999999999996	45.078	39.199	49.672	50.321000000000005	50.365	39.199	50.163999999999994	55.418	57.353	39.199	9.757	1.552	0.20500000000000002	31.843	62.743	84.78099999999999	96.86099999999999
Retrieval	MTEB CQADupstackEnglishRetrieval	29.321	39.062999999999995	40.403	40.534	35.987	44.708999999999996	45.394	45.436	35.987	44.585	49.297999999999995	51.315	35.987	8.369	1.366	0.184	29.321	54.333	74.178	86.732
Retrieval	MTEB CQADupstackGamingRetrieval	38.811	51.114000000000004	52.22	52.275000000000006	44.389	54.459	55.208999999999996	55.239000000000004	44.389	56.979	61.266	62.315	44.389	9.26	1.226	0.136	38.811	70.841	89.218	96.482
Retrieval	MTEB CQADupstackGisRetrieval	25.378	34.311	35.399	35.482	27.683999999999997	36.575	37.492	37.556	27.683999999999997	39.247	44.424	46.478	27.683999999999997	5.989	0.899	0.11199999999999999	25.378	52.195	75.764	91.012
Retrieval	MTEB CQADupstackMathematicaRetrieval	17.326	25.247000000000003	26.473000000000003	26.579000000000004	21.393	30.187	31.089	31.15	21.393	30.668	36.543	39.181	21.393	5.784000000000001	1.001	0.136	17.326	38.811	52.195	75.764

Clustering Tasks

Task	Dataset	V-Measure
Clustering	MTEB ArxivClusteringP2P	47.623282347623714
Clustering	MTEB ArxivClusteringS2S	39.86487843524932
Clustering	MTEB BiorxivClusteringP2P	39.21904455377494
Clustering	MTEB BiorxivClusteringS2S	35.43342755570654

Reranking Task

Task	Dataset	MAP	MRR
Reranking	MTEB AskUbuntuDupQuestions	62.3290291318171	75.2379853141626

STS Task

Task	Dataset	Cosine Similarity (Pearson)	Cosine Similarity (Spearman)	Euclidean (Pearson)	Euclidean (Spearman)	Manhattan (Pearson)	Manhattan (Spearman)
STS	MTEB BIOSSES	88.52002953574285	86.98752423842483	86.89442688314197	86.88631711307471	87.03723618507175	86.76041062975224

📄 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