gte-large-en-v1.5 Open-Source English Text Embedding Model - Free Support for Text Similarity and Classification Tasks

Gte Large En V1.5

Developed by Alibaba-NLP

GTE-Large is a high-performance English text embedding model that excels in multiple text similarity and classification tasks.

Text Embedding

Transformers

Supports Multiple LanguagesOpen Source License:Apache-2.0 #High-precision text embeddings #Multi-task evaluation benchmark #English semantic similarity

Downloads 891.76k

Release Time : 4/20/2024

Model Overview

This model focuses on generating high-quality sentence-level embeddings, suitable for information retrieval, text similarity calculation, and classification tasks.

Model Features

Excellent multi-task performance

Outstanding performance in various tasks (classification, clustering, retrieval, etc.) in the MTEB benchmark

High-quality sentence embeddings

Generated embeddings effectively capture semantic information, suitable for similarity calculation

Broad applicability

Supports various downstream NLP tasks, including classification, clustering, and information retrieval

Model Capabilities

Text similarity calculation

Text classification

Information retrieval

Text clustering

Sentence embedding generation

Use Cases

E-commerce

Product review sentiment analysis

Analyze sentiment tendencies in Amazon product reviews

Achieved 93.97% accuracy on the AmazonPolarity dataset

Product similarity matching

Calculate similarity between product descriptions

Customer service

Banking issue classification

Automatically classify banking customer issues

Achieved 87.33% accuracy on the Banking77 dataset

Academic research

Paper clustering

Group academic papers based on content similarity

Achieved V-measure of 48.47 on ArxivClusteringP2P

🚀 GTE-Large-EN-V1.5 Model

This README provides detailed information about the gte-large-en-v1.5 model, including its datasets, performance metrics, and license.

📚 Documentation

Model Information

Property	Details
Datasets	allenai/c4
Library Name	transformers
Tags	sentence-transformers, gte, mteb, transformers.js, sentence-similarity
License	apache-2.0
Language	en

Model Results

The model gte-large-en-v1.5 has been evaluated on multiple tasks and datasets. Here are the detailed results:

1. MTEB AmazonCounterfactualClassification (en)

Task Type: Classification
Dataset: mteb/amazon_counterfactual
Split: test
Metrics:
- Accuracy: 73.01492537313432
- AP: 35.05341696659522
- F1: 66.71270310883853

2. MTEB AmazonPolarityClassification

Task Type: Classification
Dataset: mteb/amazon_polarity
Split: test
Metrics:
- Accuracy: 93.97189999999999
- AP: 90.5952493948908
- F1: 93.95848137716877

3. MTEB AmazonReviewsClassification (en)

Task Type: Classification
Dataset: mteb/amazon_reviews_multi
Split: test
Metrics:
- Accuracy: 54.196
- F1: 53.80122334012787

4. MTEB ArguAna

Task Type: Retrieval
Dataset: mteb/arguana
Split: test
Metrics:
- MAP@1: 47.297
- MAP@10: 64.303
- MAP@100: 64.541
- MAP@1000: 64.541
- MAP@3: 60.728
- MAP@5: 63.114000000000004
- MRR@1: 48.435
- MRR@10: 64.657
- MRR@100: 64.901
- MRR@1000: 64.901
- MRR@3: 61.06
- MRR@5: 63.514
- NDCG@1: 47.297
- NDCG@10: 72.107
- NDCG@100: 72.963
- NDCG@1000: 72.963
- NDCG@3: 65.063
- NDCG@5: 69.352
- Precision@1: 47.297
- Precision@10: 9.623
- Precision@100: 0.996
- Precision@1000: 0.1
- Precision@3: 25.865
- Precision@5: 17.596
- Recall@1: 47.297
- Recall@10: 96.23
- Recall@100: 99.644
- Recall@1000: 99.644
- Recall@3: 77.596
- Recall@5: 87.98

5. MTEB ArxivClusteringP2P

Task Type: Clustering
Dataset: mteb/arxiv-clustering-p2p
Split: test
Metrics:
- V-Measure: 48.467787861077475

6. MTEB ArxivClusteringS2S

Task Type: Clustering
Dataset: mteb/arxiv-clustering-s2s
Split: test
Metrics:
- V-Measure: 43.39198391914257

7. MTEB AskUbuntuDupQuestions

Task Type: Reranking
Dataset: mteb/askubuntudupquestions-reranking
Split: test
Metrics:
- MAP: 63.12794820591384
- MRR: 75.9331442641692

8. MTEB BIOSSES

Task Type: STS
Dataset: mteb/biosses-sts
Split: test
Metrics:
- Cosine Similarity Pearson: 87.85062993863319
- Cosine Similarity Spearman: 85.39049989733459
- Euclidean Pearson: 86.00222680278333
- Euclidean Spearman: 85.45556162077396
- Manhattan Pearson: 85.88769871785621
- Manhattan Spearman: 85.11760211290839

9. MTEB Banking77Classification

Task Type: Classification
Dataset: mteb/banking77
Split: test
Metrics:
- Accuracy: 87.32792207792208
- F1: 87.29132945999555

10. MTEB BiorxivClusteringP2P

Task Type: Clustering
Dataset: mteb/biorxiv-clustering-p2p
Split: test
Metrics:
- V-Measure: 40.5779328301945

11. MTEB BiorxivClusteringS2S

Task Type: Clustering
Dataset: mteb/biorxiv-clustering-s2s
Split: test
Metrics:
- V-Measure: 37.94425623865118

12. MTEB CQADupstackAndroidRetrieval

Task Type: Retrieval
Dataset: mteb/cqadupstack-android
Split: test
Metrics:
- MAP@1: 32.978
- MAP@10: 44.45
- MAP@100: 46.19
- MAP@1000: 46.303
- MAP@3: 40.849000000000004
- MAP@5: 42.55
- MRR@1: 40.629
- MRR@10: 50.848000000000006
- MRR@100: 51.669
- MRR@1000: 51.705
- MRR@3: 47.997
- MRR@5: 49.506
- NDCG@1: 40.629
- NDCG@10: 51.102000000000004
- NDCG@100: 57.159000000000006
- NDCG@1000: 58.669000000000004
- NDCG@3: 45.738
- NDCG@5: 47.632999999999996
- Precision@1: 40.629
- Precision@10: 9.700000000000001
- Precision@100: 1.5970000000000002
- Precision@1000: 0.202
- Precision@3: 21.698
- Precision@5: 15.393
- Recall@1: 32.978
- Recall@10: 63.711
- Recall@100: 88.39399999999999
- Recall@1000: 97.513
- Recall@3: 48.025
- Recall@5: 53.52

13. MTEB CQADupstackEnglishRetrieval

Task Type: Retrieval
Dataset: mteb/cqadupstack-english
Split: test
Metrics:
- MAP@1: 30.767
- MAP@10: 42.195
- MAP@100: 43.541999999999994
- MAP@1000: 43.673
- MAP@3: 38.561
- MAP@5: 40.532000000000004
- MRR@1: 38.79
- MRR@10: 48.021
- MRR@100: 48.735
- MRR@1000: 48.776
- MRR@3: 45.594
- MRR@5: 46.986
- NDCG@1: 38.79
- NDCG@10: 48.468
- NDCG@100: 53.037
- NDCG@1000: 55.001999999999995
- NDCG@3: 43.409
- NDCG@5: 45.654
- Precision@1: 38.79
- Precision@10: 9.452
- Precision@100: 1.518
- Precision@1000: 0.201
- Precision@3: 21.21
- Precision@5: 15.171999999999999
- Recall@1: 30.767
- Recall@10: 60.118
- Recall@100: 79.271
- Recall@1000: 91.43299999999999
- Recall@3: 45.36
- Recall@5: 51.705

14. MTEB CQADupstackGamingRetrieval

Task Type: Retrieval
Dataset: mteb/cqadupstack-gaming
Split: test
Metrics:
- MAP@1: 40.007
- MAP@10: 53.529
- MAP@100: 54.602
- MAP@1000: 54.647
- MAP@3: 49.951
- MAP@5: 52.066
- MRR@1: 45.705
- MRR@10: 56.745000000000005
- MRR@100: 57.43899999999999
- MRR@1000: 57.462999999999994
- MRR@3: 54.25299999999999
- MRR@5: 55.842000000000006
- NDCG@1: 45.705
- NDCG@10: 59.809
- NDCG@100: 63.837999999999994
- NDCG@1000: 64.729
- NDCG@3: 53.994
- NDCG@5: 57.028
- Precision@1: 45.705
- Precision@10: 9.762
- Precision@100: 1.275
- Precision@1000: 0.13899999999999998
- Precision@3: 24.368000000000002
- Precision@5: 16.84
- Recall@1: 40.007
- Recall@10: 75.017
- Recall@100: 91.99000000000001
- Recall@1000: 98.265
- Recall@3: 59.704
- Recall@5: 67.109

15. MTEB CQADupstackGisRetrieval

Task Type: Retrieval
Dataset: mteb/cqadupstack-gis
Split: test
Metrics:
- MAP@1: 26.639000000000003
- MAP@10: 35.926
- MAP@100: 37.126999999999995
- MAP@1000: 37.202
- MAP@3: 32.989000000000004
- MAP@5: 34.465
- MRR@1: 28.475
- MRR@10: 37.7
- MRR@100: 38.753
- MRR@1000: 38.807
- MRR@3: 35.066
- MRR@5: 36.512
- NDCG@1: 28.475
- NDCG@10: 41.245
- NDCG@100: 46.814
- NDCG@1000: 48.571
- NDCG@3: 35.528999999999996
- NDCG@5: 38.066
- Precision@1: 28.475
- Precision@10: 6.497
- Precision@100: 0.9650000000000001
- Precision@1000: 0.11499999999999999
- Precision@3: 15.065999999999999
- Precision@5: 10.599
- Recall@1: 26.639000000000003
- Recall@10: 55.759
- Recall@100: 80.913
- Recall@1000: 93.929
- Recall@3: 40.454
- Recall@5: 46.439

16. MTEB CQADupstackMathematicaRetrieval

Task Type: Retrieval
Dataset: mteb/cqadupstack-mathematica
Split: test
Metrics:
- MAP@1: 15.767999999999999
- MAP@10: 24.811
- MAP@100: 26.064999999999998
- MAP@1000: 26.186999999999998
- MAP@3: 21.736
- MAP@5: 23.283
- MRR@1: 19.527
- MRR@10: 29.179
- MRR@100: 30.153999999999996
- MRR@1000: 30.215999999999998
- MRR@3: 26.223000000000003
- MRR@5: 27.733999999999998
- NDCG@1: 19.527
- NDCG@10: 30.786
- NDCG@100: 36.644
- NDCG@1000: 39.440999999999995
- NDCG@3: 24.958
- NDCG@5: 27.392
- Precision@1: 19.527
- Precision@10: 5.995

📄 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