gte-Qwen2-1.5B-instruct Open-Source Text Embedding Model - Free Deployment, Support for Multilingual Long-Text Processing

Home

Gte Qwen2 1.5B Instruct

Developed by Alibaba-NLP

A general-purpose text embedding model based on Qwen2-1.5B, supporting multilingual and long-text processing

Text Embedding

Transformers

Open Source License:Apache-2.0 #Long-Text Embedding #Multilingual Understanding #Instruction Fine-Tuning

Downloads 242.12k

Release Time : 6/29/2024

Model Overview

This model is the latest addition to the General Text Embedding (gte) series, focusing on generating high-quality text embeddings suitable for various NLP tasks such as information retrieval and semantic similarity calculation.

Model Features

Bidirectional Attention Mechanism

Enhances contextual understanding and improves embedding quality

Efficient Instruction Fine-Tuning

Only fine-tunes the query side for improved efficiency

Multilingual Support

Trained on cross-domain, multi-scenario multilingual text corpora

Long-Text Processing

Supports input lengths of up to 32k tokens

Model Capabilities

Text Embedding Generation

Semantic Similarity Calculation

Information Retrieval

Multilingual Text Processing

Use Cases

Information Retrieval

Web Search Query

Retrieves relevant document passages based on user queries

Achieved a score of 67.16 in MTEB evaluation

Semantic Similarity

Document Similarity Calculation

Computes semantic similarity between different documents

🚀 gte-qwen2-7B-instruct

This is a model related to sentence - similarity tasks, leveraging technologies such as mteb, sentence - transformers, transformers, and Qwen2.

📚 Documentation

Model Information

Property	Details
Model Name	gte - qwen2 - 7B - instruct
Tags	mteb, sentence - transformers, transformers, Qwen2, sentence - similarity

Performance Results

The model has been tested on multiple datasets, and the following are the performance metrics:

1. MTEB AmazonCounterfactualClassification (en)

Task Type: Classification
Dataset Revision: e8379541af4e31359cca9fbcf4b00f2671dba205
Metrics:
- Accuracy: 83.98507462686567
- AP: 50.93015252587014
- F1: 78.50416599051215

2. MTEB AmazonPolarityClassification

Task Type: Classification
Dataset Revision: e2d317d38cd51312af73b3d32a06d1a08b442046
Metrics:
- Accuracy: 96.61065
- AP: 94.89174052954196
- F1: 96.60942596940565

3. MTEB AmazonReviewsClassification (en)

Task Type: Classification
Dataset Revision: 1399c76144fd37290681b995c656ef9b2e06e26d
Metrics:
- Accuracy: 55.614000000000004
- F1: 54.90553480294904

4. MTEB ArguAna

Task Type: Retrieval
Dataset Revision: c22ab2a51041ffd869aaddef7af8d8215647e41a
Metrics:
- MAP@1: 45.164
- MAP@10: 61.519
- MAP@100: 61.769
- MAP@1000: 61.769
- MRR@1: 46.088
- MRR@10: 61.861
- MRR@100: 62.117999999999995
- MRR@1000: 62.117999999999995
- NDCG@1: 45.164
- NDCG@10: 69.72
- NDCG@100: 70.719
- NDCG@1000: 70.719
- Precision@1: 45.164
- Precision@10: 9.545
- Precision@100: 0.996
- Precision@1000: 0.1
- Recall@1: 45.164
- Recall@10: 95.448
- Recall@100: 99.644
- Recall@1000: 99.644

5. MTEB ArxivClusteringP2P

Task Type: Clustering
Dataset Revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d
Metrics:
- V - measure: 50.511868162026175

6. MTEB ArxivClusteringS2S

Task Type: Clustering
Dataset Revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53
Metrics:
- V - measure: 45.007803189284004

7. MTEB AskUbuntuDupQuestions

Task Type: Reranking
Dataset Revision: 2000358ca161889fa9c082cb41daa8dcfb161a54
Metrics:
- MAP: 64.55292107723382
- MRR: 77.66158818097877

8. MTEB BIOSSES

Task Type: STS
Dataset Revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
Metrics:
- Cosine Similarity Pearson: 85.65459047085452
- Cosine Similarity Spearman: 82.10729255710761
- Euclidean Pearson: 82.78079159312476
- Euclidean Spearman: 80.50002701880933
- Manhattan Pearson: 82.41372641383016
- Manhattan Spearman: 80.57412509272639

9. MTEB Banking77Classification

Task Type: Classification
Dataset Revision: 0fd18e25b25c072e09e0d92ab615fda904d66300
Metrics:
- Accuracy: 87.30844155844156
- F1: 87.25307322443255

10. MTEB BiorxivClusteringP2P

Task Type: Clustering
Dataset Revision: 65b79d1d13f80053f67aca9498d9402c2d9f1f40
Metrics:
- V - measure: 43.20754608934859

11. MTEB BiorxivClusteringS2S

Task Type: Clustering
Dataset Revision: 258694dd0231531bc1fd9de6ceb52a0853c6d908
Metrics:
- V - measure: 38.818037697335505

12. MTEB CQADupstackAndroidRetrieval

Task Type: Retrieval
Dataset Revision: f46a197baaae43b4f621051089b82a364682dfeb
Metrics:
- MAP@1: 35.423
- MAP@10: 47.198
- MAP@100: 48.899
- MAP@1000: 49.004
- MRR@1: 42.918
- MRR@10: 53.299
- MRR@100: 54.032000000000004
- MRR@1000: 54.055
- NDCG@1: 42.918
- NDCG@10: 53.98
- NDCG@100: 59.57
- NDCG@1000: 60.879000000000005
- Precision@1: 42.918
- Precision@10: 10.299999999999999
- Precision@100: 1.687
- Precision@1000: 0.211
- Recall@1: 35.423
- Recall@10: 66.824
- Recall@100: 89.564
- Recall@1000: 97.501

13. MTEB CQADupstackEnglishRetrieval

Task Type: Retrieval
Dataset Revision: ad9991cb51e31e31e430383c75ffb2885547b5f0
Metrics:
- MAP@1: 33.205
- MAP@10: 44.859
- MAP@100: 46.135
- MAP@1000: 46.259
- MRR@1: 41.146
- MRR@10: 50.621
- MRR@100: 51.207
- MRR@1000: 51.246
- NDCG@1: 41.146
- NDCG@10: 50.683
- NDCG@100: 54.82
- NDCG@1000: 56.69
- Precision@1: 41.146
- Precision@10: 9.439
- Precision@100: 1.465
- Precision@1000: 0.194
- Recall@1: 33.205
- Recall@10: 61.028999999999996
- Recall@100: 78.152
- Recall@1000: 89.59700000000001

14. MTEB CQADupstackGamingRetrieval

Task Type: Retrieval
Dataset Revision: 4885aa143210c98657558c04aaf3dc47cfb54340
Metrics:
- MAP@1: 41.637
- MAP@10: 55.162
- MAP@100: 56.142
- MAP@1000: 56.188
- MRR@1: 47.524
- MRR@10: 58.243
- MRR@100: 58.879999999999995
- MRR@1000: 58.9
- NDCG@1: 47.524
- NDCG@10: 61.305
- NDCG@100: 65.077
- NDCG@1000: 65.941
- Precision@1: 47.524
- Precision@10: 9.918000000000001
- Precision@100: 1.276
- Precision@1000: 0.13899999999999998
- Recall@1: 41.637
- Recall@10: 76.185
- Recall@100: 92.149
- Recall@1000: 98.199

15. MTEB CQADupstackGisRetrieval

Task Type: Retrieval
Dataset Revision: 5003b3064772da1887988e05400cf3806fe491f2
Metrics:
- MAP@1: 26.27
- MAP@10: 37.463
- MAP@100: 38.434000000000005
- MAP@1000: 38.509
- MRR@1: 28.588
- MRR@10: 39.383
- MRR@100: 40.23
- MRR@1000: 40.281
- NDCG@1: 28.588
- NDCG@10: 43.511
- NDCG@100: 48.274
- NDCG@1000: 49.975
- Precision@1: 28.588
- Precision@10: 6.893000000000001
- Precision@100: 0.9900000000000001
- Precision@1000: 0.117
- Recall@1: 26.27
- Recall@10: 60.284000000000006
- Recall@100: 81.902
- Recall@1000: 94.43

16. MTEB CQADupstackMathematicaRetrieval

Task Type: Retrieval
Dataset Revision: 90fceea13679c63fe563ded68f3b6f06e50061de
Metrics:
- MAP@1: 18.168
- MAP@10: 28.410000000000004
- MAP@100: 29.78
- MAP@1000: 29.892999999999997
- MRR@1: 23.507
- MRR@10: 33.382
- MRR@100: 34.404
- MRR@1000: 34.467999999999996
- NDCG@1: 23.507
- NDCG@10: 34.571000000000005
- NDCG@100: 40.663
- NDCG@1000: 43.236000000000004
- Precision@1: 23.507
- Precision@10: 6.654
- Precision@100: 1.113
- Precision@1000: 0.146
- Recall@1: 18.168
- Recall@10: 37.88
- Recall@100: 68.04
- Recall@1000: 92.18

📄 License

This project is licensed under the Apache 2.0 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