Snowflake Arctic Embed XS Open-Source Model - Achieve Sentence Similarity and Feature Extraction with a Lightweight Tool

Snowflake Arctic Embed Xs

Developed by Snowflake

Snowflake Arctic Embed XS is a lightweight sentence embedding model focused on sentence similarity and feature extraction tasks.

Text Embedding

Transformers

Downloads 125.31k

Release Time : 4/12/2024

Model Overview

This model is based on the sentence-transformers architecture, primarily used for generating sentence embeddings, suitable for tasks like sentence similarity calculation, text classification, and clustering.

Model Features

Lightweight design

The XS version is optimized for efficiency and performance, suitable for resource-constrained environments.

Multi-task support

Supports various NLP tasks including sentence similarity, classification, clustering, and retrieval.

High performance

Performs excellently in multiple MTEB benchmarks, particularly in classification and retrieval tasks.

Model Capabilities

Sentence similarity calculation

Text feature extraction

Text classification

Text clustering

Information retrieval

Use Cases

E-commerce

Product review classification

Sentiment polarity classification for Amazon product reviews

Achieved 70.04% accuracy in the MTEB Amazon Polarity classification task

Counterfactual detection

Identifying counterfactual statements in Amazon product descriptions

Achieved 65.09% accuracy in the MTEB AmazonCounterfactual classification task

Technical support

Duplicate question detection

Identifying duplicate technical questions in AskUbuntu forums

Achieved 60.25 average precision in the MTEB AskUbuntu duplicate questions task

Academic research

Paper clustering

Topic clustering for arXiv and bioRxiv academic papers

Achieved 43.50 V-measure in the arXiv P2P clustering task

🚀 snowflake-snowflake-arctic-embed-xs

This model is designed for sentence - similarity tasks and provides feature - extraction capabilities. It has been tested on multiple datasets in the MTEB benchmark, showing its performance in various NLP tasks such as classification, retrieval, clustering, reranking, and STS.

📚 Documentation

Model Information

Property	Details
Model Type	Sentence Transformer for sentence - similarity and feature - extraction
Tags	sentence - transformers, feature - extraction, sentence - similarity, mteb, arctic, snowflake - arctic - embed, transformers.js

Results on Datasets

1. MTEB AmazonCounterfactualClassification (en)

Task: Classification
Metrics:
- Accuracy: 65.08955223880598
- AP: 28.514291209445364
- F1: 59.2604580112738

2. MTEB AmazonPolarityClassification

Task: Classification
Metrics:
- Accuracy: 70.035375
- AP: 64.29444264250405
- F1: 69.78382333907138

3. MTEB AmazonReviewsClassification (en)

Task: Classification
Metrics:
- Accuracy: 35.343999999999994
- F1: 34.69618251902858

4. MTEB ArguAna

Task: Retrieval
Metrics:
- MAP@1: 28.592000000000002
- MAP@10: 43.597
- MAP@100: 44.614
- MAP@1000: 44.624
- MAP@3: 38.928000000000004
- MAP@5: 41.453
- MRR@1: 29.232000000000003
- MRR@10: 43.829
- MRR@100: 44.852
- MRR@1000: 44.862
- MRR@3: 39.118
- MRR@5: 41.703
- NDCG@1: 28.592000000000002
- NDCG@10: 52.081
- NDCG@100: 56.37
- NDCG@1000: 56.598000000000006
- NDCG@3: 42.42
- NDCG@5: 46.965
- Precision@1: 28.592000000000002
- Precision@10: 7.922999999999999
- Precision@100: 0.979
- Precision@1000: 0.1
- Precision@3: 17.52
- Precision@5: 12.717
- Recall@1: 28.592000000000002
- Recall@10: 79.232
- Recall@100: 97.866
- Recall@1000: 99.57300000000001
- Recall@3: 52.559999999999995
- Recall@5: 63.585

5. MTEB ArxivClusteringP2P

Task: Clustering
Metrics:
- V - measure: 43.50220588953974

6. MTEB ArxivClusteringS2S

Task: Clustering
Metrics:
- V - measure: 32.08725826118282

7. MTEB AskUbuntuDupQuestions

Task: Reranking
Metrics:
- MAP: 60.25381587694928
- MRR: 73.79776194873148

8. MTEB BIOSSES

Task: STS
Metrics:
- Cosine Similarity Pearson: 85.47489332445278
- Cosine Similarity Spearman: 84.05432487336698
- Euclidean Pearson: 84.5108222177219
- Euclidean Spearman: 84.05432487336698
- Manhattan Pearson: 84.20440618321464
- Manhattan Spearman: 83.9290208134097

9. MTEB Banking77Classification

Task: Classification
Metrics:
- Accuracy: 76.37337662337663
- F1: 75.33296834885043

10. MTEB BigPatentClustering

Task: Clustering
Metrics:
- V - measure: 21.31174373264835

11. MTEB BiorxivClusteringP2P

Task: Clustering
Metrics:
- V - measure: 34.481973521597844

12. MTEB BiorxivClusteringS2S

Task: Clustering
Metrics:
- V - measure: 26.14094256567341

13. MTEB CQADupstackAndroidRetrieval

Task: Retrieval
Metrics:
- MAP@1: 32.527
- MAP@10: 43.699
- MAP@100: 45.03
- MAP@1000: 45.157000000000004
- MAP@3: 39.943
- MAP@5: 42.324
- MRR@1: 39.771
- MRR@10: 49.277
- MRR@100: 49.956
- MRR@1000: 50.005
- MRR@3: 46.304
- MRR@5: 48.493
- NDCG@1: 39.771
- NDCG@10: 49.957
- NDCG@100: 54.678000000000004
- NDCG@1000: 56.751
- NDCG@3: 44.608
- NDCG@5: 47.687000000000005
- Precision@1: 39.771
- Precision@10: 9.557
- Precision@100: 1.5010000000000001
- Precision@1000: 0.194
- Precision@3: 21.173000000000002
- Precision@5: 15.794
- Recall@1: 32.527
- Recall@10: 61.791
- Recall@100: 81.49300000000001
- Recall@1000: 95.014
- Recall@3: 46.605000000000004
- Recall@5: 54.83

14. MTEB CQADupstackEnglishRetrieval

Task: Retrieval
Metrics:
- MAP@1: 29.424
- MAP@10: 38.667
- MAP@100: 39.771
- MAP@1000: 39.899
- MAP@3: 35.91
- MAP@5: 37.45
- MRR@1: 36.687999999999995
- MRR@10: 44.673
- MRR@100: 45.289
- MRR@1000: 45.338
- MRR@3: 42.601
- MRR@5: 43.875
- NDCG@1: 36.687999999999995
- NDCG@10: 44.013000000000005
- NDCG@100: 48.13
- NDCG@1000: 50.294000000000004
- NDCG@3: 40.056999999999995
- NDCG@5: 41.902
- Precision@1: 36.687999999999995
- Precision@10: 8.158999999999999
- Precision@100: 1.321
- Precision@1000: 0.179
- Precision@3: 19.045
- Precision@5: 13.427
- Recall@1: 29.424
- Recall@10: 53.08500000000001
- Recall@100: 70.679
- Recall@1000: 84.66
- Recall@3: 41.399
- Recall@5: 46.632

15. MTEB CQADupstackGamingRetrieval

Task: Retrieval
Metrics:
- MAP@1: 39.747
- MAP@10: 51.452
- MAP@100: 52.384
- MAP@1000: 52.437
- MAP@3: 48.213
- MAP@5: 50.195
- MRR@1: 45.391999999999996
- MRR@10: 54.928
- MRR@100: 55.532000000000004
- MRR@1000: 55.565
- MRR@3: 52.456
- MRR@5: 54.054
- NDCG@1: 45.391999999999996
- NDCG@10: 57.055
- NDCG@100: 60.751999999999995
- NDCG@1000: 61.864
- NDCG@3: 51.662
- NDCG@5: 54.613
- Precision@1: 45.391999999999996
- Precision@10: 9.103
- Precision@100: 1.1780000000000002
- Precision@1000: 0.132
- Precision@3: 22.717000000000002
- Precision@5: 15.812000000000001
- Recall@1: 39.747
- Recall@10: 70.10499999999999
- Recall@100: 86.23100000000001
- Recall@1000: 94.025
- Recall@3: 55.899
- Recall@5: 63.05500000000001

16. MTEB CQADupstackGisRetrieval

Task: Retrieval
Metrics:
- MAP@1: 27.168999999999997
- MAP@10: 34.975
- MAP@100: 35.94
- MAP@1000: 36.021
- MAP@3: 32.35
- MAP@5: 33.831
- MRR@1: 28.701
- MRR@10: 36.698
- MRR@100: 37.546
- MRR@1000: 37.613
- MRR@3: 34.256
- MRR@5: 35.685
- NDCG@1: 28.701
- NDCG@10: 39.639
- NDCG@100: 44.389
- NDCG@1000: 46.46
- NDCG@3: 34.52
- NDCG@5: 37.076
- Precision@1: 28.701
- Precision@10: 5.955
- Precision@100: 0.8880000000000001
- Precision@1000: 0.109
- Precision@3: 14.274999999999999
- Precision@5: 10.011000000000001
- Recall@1: 27.168999999999997
- Recall@10: 52.347
- Recall@100: 74.1
- Recall@1000: 89.739
- Recall@3: 38.567
- Recall@5: 44.767

17. MTEB CQADupstackMathematicaRetrieval

Task: Retrieval
Metrics:
- MAP@1: 15.872
- MAP@10: 23.153000000000002
- MAP@100: 24.311
- MAP@1000: 24.432000000000002
- MAP@3: 20.707
- MAP@5: 21.921
- MRR@1: 19.776
- MRR@10: 27.755999999999997
- MRR@100: 28.709
- MRR@1000: 28.778
- MRR@3: 25.186999999999998
- MRR@5: 26. (The data seems incomplete here)

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