multilingual-e5-large-instruct-GGUF Open Source Model - Supports Text Embedding and Classification Tasks in Over 100 Languages

Multilingual E5 Large Instruct GGUF

Developed by KeyurRamoliya

Multilingual E5 large instruction model supporting text embedding and classification tasks in over 100 languages

Large Language Model Supports Multiple LanguagesOpen Source License:MIT #Multilingual Embedding #Instruction Fine-tuning #Text Classification

Downloads 224

Release Time : 8/23/2024

Model Overview

This is a multilingual text embedding model based on the E5 architecture, specifically optimized for instruction-following tasks. It supports a wide range of languages and is suitable for various natural language processing tasks such as classification, retrieval, and clustering.

Model Features

Multilingual Support

Supports text processing in over 100 languages, including major languages and many minority languages

Instruction Optimization

Specifically optimized for instruction-following tasks, enabling better understanding and execution of user instructions

High-performance Classification

Demonstrates excellent text classification capabilities in MTEB benchmarks, achieving 96.29% accuracy in English classification

Powerful Retrieval Capability

Outstanding performance in ArguAna retrieval tasks with an average precision@10 of 49.221

Model Capabilities

Text Embedding

Multilingual Text Processing

Text Classification

Information Retrieval

Text Clustering

Instruction Understanding

Use Cases

E-commerce

Multilingual Product Review Classification

Sentiment classification of multilingual product reviews on platforms like Amazon

Achieves 96.29% accuracy in English review classification

Counterfactual Review Detection

Identifying counterfactual reviews on e-commerce platforms

Achieves 76.24% accuracy in English counterfactual classification tasks

Information Retrieval

Argument Retrieval

Retrieving relevant arguments in debate datasets

Achieves average precision@10 of 49.221 in ArguAna tasks

Academic Research

Paper Clustering

Clustering arXiv papers with V-measure reaching 46.40

V-measure of 46.40 in arXiv paper clustering tasks

🚀 Multilingual-e5-large-instruct

This is a multilingual model that provides performance metrics across various tasks and datasets, including classification, retrieval, clustering, reranking, STS, and bitext mining.

📚 Documentation

Model Information

Property	Details
Base Model	intfloat/multilingual-e5-large-instruct
Language	Multilingual, including af, am, ar, as, az, be, bg, bn, br, bs, ca, cs, cy, da, de, el, en, eo, es, et, eu, fa, fi, fr, fy, ga, gd, gl, gu, ha, he, hi, hr, hu, hy, id, is, it, ja, jv, ka, kk, km, kn, ko, ku, ky, la, lo, lt, lv, mg, mk, ml, mn, mr, ms, my, ne, nl, 'no', om, or, pa, pl, ps, pt, ro, ru, sa, sd, si, sk, sl, so, sq, sr, su, sv, sw, ta, te, th, tl, tr, ug, uk, ur, uz, vi, xh, yi, zh
License	MIT
Tags	mteb, sentence-transformers, transformers, llama-cpp, gguf-my-repo

Model Performance Results

Classification Tasks

MTEB AmazonCounterfactualClassification (en):
- Accuracy: 76.23880597014924
- AP: 39.07351965022687
- F1: 70.04836733862683
MTEB AmazonCounterfactualClassification (de):
- Accuracy: 66.71306209850107
- AP: 79.01499914759529
- F1: 64.81951817560703
MTEB AmazonCounterfactualClassification (en-ext):
- Accuracy: 73.85307346326837
- AP: 22.447519885878737
- F1: 61.0162730745633
MTEB AmazonCounterfactualClassification (ja):
- Accuracy: 76.04925053533191
- AP: 23.44983217128922
- F1: 62.5723230907759
MTEB AmazonPolarityClassification:
- Accuracy: 96.28742500000001
- AP: 94.8449918887462
- F1: 96.28680923610432
MTEB AmazonReviewsClassification (en):
- Accuracy: 56.716
- F1: 55.76510398266401
MTEB AmazonReviewsClassification (de):
- Accuracy: 52.99999999999999
- F1: 52.00829994765178
MTEB AmazonReviewsClassification (es):
- Accuracy: 48.806000000000004
- F1: 48.082345914983634
MTEB AmazonReviewsClassification (fr):
- Accuracy: 48.507999999999996
- F1: 47.68752844642045
MTEB AmazonReviewsClassification (ja):
- Accuracy: 47.709999999999994
- F1: 47.05870376637181
MTEB AmazonReviewsClassification (zh):
- Accuracy: 44.662000000000006
- F1: 43.42371965372771
MTEB Banking77Classification:
- Accuracy: 85.73376623376623
- F1: 85.68480707214599

Retrieval Tasks

MTEB ArguAna:
- MAP@1: 31.721
- MAP@10: 49.221
- MAP@100: 49.884
- MAP@1000: 49.888
- MAP@3: 44.31
- MAP@5: 47.276
- MRR@1: 32.432
- MRR@10: 49.5
- MRR@100: 50.163000000000004
- MRR@1000: 50.166
- MRR@3: 44.618
- MRR@5: 47.541
- NDCG@1: 31.721
- NDCG@10: 58.384
- NDCG@100: 61.111000000000004
- NDCG@1000: 61.187999999999995
- NDCG@3: 48.386
- NDCG@5: 53.708999999999996
- Precision@1: 31.721
- Precision@10: 8.741
- Precision@100: 0.991
- Precision@1000: 0.1
- Precision@3: 20.057
- Precision@5: 14.609
- Recall@1: 31.721
- Recall@10: 87.411
- Recall@100: 99.075
- Recall@1000: 99.644
- Recall@3: 60.171
- Recall@5: 73.044
MTEB CQADupstackRetrieval:
- MAP@1: 27.764166666666668
- MAP@10: 37.298166666666674
- MAP@100: 38.530166666666666
- MAP@1000: 38.64416666666667
- MAP@3: 34.484833333333334
- MAP@5: 36.0385
- MRR@1: 32.93558333333333
- MRR@10: 41.589749999999995
- MRR@100: 42.425333333333334
- MRR@1000: 42.476333333333336
- MRR@3: 39.26825
- MRR@5: 40.567083333333336
- NDCG@1: 32.93558333333333
- NDCG@10: 42.706583333333334
- NDCG@100: 47.82483333333333
- NDCG@1000: 49.95733333333334
- NDCG@3: 38.064750000000004
- NDCG@5: 40.18158333333333
- Precision@1: 32.93558333333333
- Precision@10: 7.459833333333334
- Precision@100: 1.1830833333333335
- Precision@1000: 0.15608333333333332
- Precision@3: 17.5235
- Precision@5: 12.349833333333333
- Recall@1: 27.764166666666668
- Recall@10: 54.31775
- Recall@100: 76.74350000000001
- Recall@1000: 91.45208333333332
- Recall@3: 41.23425
- Recall@5: 46.73983333333334
MTEB ClimateFEVER:
- MAP@1: 12.969
- MAP@10: 21.584999999999997
- MAP@100: 23.3
- MAP@1000: 23.5
- MAP@3: 18.218999999999998
- MAP@5: 19.983
- MRR@1: 29.316
- MRR@10: 40.033
- MRR@100: 40.96
- MRR@1000: 41.001
- MRR@3: 37.123
- MRR@5: 38.757999999999996
- NDCG@1: 29.316
- NDCG@10: 29.858
- NDCG@100: 36.756
- NDCG@1000: 40.245999999999995

Clustering Tasks

MTEB ArxivClusteringP2P:
- V-measure: 46.40419580759799
MTEB ArxivClusteringS2S:
- V-measure: 40.48593255007969
MTEB BiorxivClusteringP2P:
- V-measure: 40.935218072113855
MTEB BiorxivClusteringS2S:
- V-measure: 36.276389017675264

Reranking Tasks

MTEB AskUbuntuDupQuestions:
- MAP: 63.889179122289995
- MRR: 77.61146286769556

STS Tasks

MTEB BIOSSES:
- Cosine Similarity Pearson: 88.15075203727929
- Cosine Similarity Spearman: 86.9622224570873
- Euclidean Pearson: 86.70473853624121
- Euclidean Spearman: 86.9622224570873
- Manhattan Pearson: 86.21089380980065
- Manhattan Spearman: 86.75318154937008

BitextMining Tasks

MTEB BUCC (de-en):
- Accuracy: 99.65553235908142
- F1: 99.60681976339595
- Precision: 99.58246346555325
- Recall: 99.65553235908142
MTEB BUCC (fr-en):
- Accuracy: 99.26260180497468
- F1: 99.14520507740848
- Precision: 99.08650671362535
- Recall: 99.26260180497468
MTEB BUCC (ru-en):
- Accuracy: 98.07412538967787
- F1: 97.86629719431936
- Precision: 97.76238309664012
- Recall: 98.07412538967787
MTEB BUCC (zh-en):
- Accuracy: 99.42074776197998
- F1: 99.38564156573635
- Precision: 99.36808846761454
- Recall: 99.42074776197998

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