multilingual e5 large

提供商intfloat
分类feature-extraction
许可证mit
下载量4.8K
星标1

简介

Multilingual E5 Large 是一款由 intfloat 开发的开源多语言文本嵌入模型。与直接生成文本的 LLM 不同，它专注于将文本转化为高质量的向量（Embedding），是构建 RAG（检索增强生成）系统的核心组件。该模型在多语言语义对齐上表现出色，能够将中文、英文等多种语言映射到同一向量空间，这意味着你可以用英文查询中文文档并获得精准匹配。对于开发者而言，它上手简单，可直接通过 Hugging Face 或本地向量数据库部署，是替代闭源 Embedding API、降低检索成本的理想选择。

核心亮点

  • 强大的多语言语义对齐，支持跨语言检索
  • RAG 架构的核心组件，大幅提升知识库召回率
  • MIT 协议开源，可本地化部署，无 API 调用成本
  • 语义表征能力强，适用于文本聚类与相似度计算

使用方法

安装依赖
# 安装 Hugging Face transformers
pip install transformers torch
SDK 使用
# 使用 transformers 加载模型
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("intfloat/multilingual-e5-large")
tokenizer = AutoTokenizer.from_pretrained("intfloat/multilingual-e5-large")

Hugging Face 下载

我们推荐使用命令行或者 Hugging Face Hub SDK 来进行模型的下载。

操作指引：在下载前，请先通过如下命令安装 huggingface_hub：

操作指引 
pip install -U huggingface_hub

命令行下载

下载完整模型库

下载完整模型库 
huggingface-cli download intfloat/multilingual-e5-large

下载单个文件到指定本地文件夹（以下载 config.json 到当前路径下 ./dir 目录为例）

下载单个文件到指定本地文件夹（以下载 config.json 到当前路径下 ./dir 目录为例） 
huggingface-cli download intfloat/multilingual-e5-large config.json --local-dir ./dir

更多命令行下载选项，可参见官方文档

SDK 下载

SDK 下载 
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('intfloat/multilingual-e5-large')

Git 下载

请确保 lfs 已经被正确安装

Git 下载 
git lfs install
git clone https://huggingface.co/intfloat/multilingual-e5-large

如果您希望跳过 lfs 大文件下载，可以使用如下命令

跳过 LFS 
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/intfloat/multilingual-e5-large

模型文件托管在 Hugging Face Hub，使用 HF CLI / SDK / Git 直接下载，不经过本站。

PyTorch / Transformers 使用

安装 Transformers

安装 Transformers 
pip install -U transformers torch

模型加载和推理

模型加载和推理 
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('intfloat/multilingual-e5-large')
tokenizer = AutoTokenizer.from_pretrained('intfloat/multilingual-e5-large')

模型下载

我们推荐使用命令行或者 ModelScope SDK 来进行模型的下载。

操作指引：在下载前，请先通过如下命令安装 ModelScope：

操作指引 
pip install modelscope

命令行下载

下载完整模型库

下载完整模型库 
modelscope download --model intfloat/multilingual-e5-large

下载单个文件到指定本地文件夹（以下载 README.md 到当前路径下 dir 目录为例）

下载单个文件到指定本地文件夹（以下载 README.md 到当前路径下 dir 目录为例） 
modelscope download --model intfloat/multilingual-e5-large README.md --local_dir ./dir

更多更丰富的命令行下载选项，可参见具体文档

SDK 下载

SDK 下载 
# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('intfloat/multilingual-e5-large')

Git 下载

请确保 lfs 已经被正确安装

Git 下载 
git lfs install
git clone https://www.modelscope.cn/intfloat/multilingual-e5-large.git

如果您希望跳过 lfs 大文件下载，可以使用如下命令

跳过 LFS 
GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/intfloat/multilingual-e5-large.git

ModelScope 模型页直接下载模型文件；无需将模型文件放在本站服务器。

Notebook 快速开发

下载并安装 ModelScope library

下载并安装 ModelScope library 
pip install "modelscope[audio,cv,nlp,multi-modal,science]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

模型加载和推理

模型加载和推理 
from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks

p = pipeline('text-generation', 'intfloat/multilingual-e5-large')

完整文档

来源: HuggingFace

---
tags:

  • mteb

  • Sentence Transformers

  • sentence-similarity

  • feature-extraction

  • sentence-transformers

model-index:
  • name: multilingual-e5-large

results:
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (en)
config: en
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 79.05970149253731
- type: ap
value: 43.486574390835635
- type: f1
value: 73.32700092140148
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (de)
config: de
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 71.22055674518201
- type: ap
value: 81.55756710830498
- type: f1
value: 69.28271787752661
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (en-ext)
config: en-ext
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 80.41979010494754
- type: ap
value: 29.34879922376344
- type: f1
value: 67.62475449011278
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (ja)
config: ja
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 77.8372591006424
- type: ap
value: 26.557560591210738
- type: f1
value: 64.96619417368707
- task:
type: Classification
dataset:
type: mteb/amazon_polarity
name: MTEB AmazonPolarityClassification
config: default
split: test
revision: e2d317d38cd51312af73b3d32a06d1a08b442046
metrics:
- type: accuracy
value: 93.489875
- type: ap
value: 90.98758636917603
- type: f1
value: 93.48554819717332
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (en)
config: en
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 47.564
- type: f1
value: 46.75122173518047
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (de)
config: de
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 45.400000000000006
- type: f1
value: 44.17195682400632
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (es)
config: es
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 43.068
- type: f1
value: 42.38155696855596
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (fr)
config: fr
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 41.89
- type: f1
value: 40.84407321682663
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (ja)
config: ja
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 40.120000000000005
- type: f1
value: 39.522976223819114
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (zh)
config: zh
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 38.832
- type: f1
value: 38.0392533394713
- task:
type: Retrieval
dataset:
type: arguana
name: MTEB ArguAna
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 30.725
- type: map_at_10
value: 46.055
- type: map_at_100
value: 46.900999999999996
- type: map_at_1000
value: 46.911
- type: map_at_3
value: 41.548
- type: map_at_5
value: 44.297
- type: mrr_at_1
value: 31.152
- type: mrr_at_10
value: 46.231
- type: mrr_at_100
value: 47.07
- type: mrr_at_1000
value: 47.08
- type: mrr_at_3
value: 41.738
- type: mrr_at_5
value: 44.468999999999994
- type: ndcg_at_1
value: 30.725
- type: ndcg_at_10
value: 54.379999999999995
- type: ndcg_at_100
value: 58.138
- type: ndcg_at_1000
value: 58.389
- type: ndcg_at_3
value: 45.156
- type: ndcg_at_5
value: 50.123
- type: precision_at_1
value: 30.725
- type: precision_at_10
value: 8.087
- type: precision_at_100
value: 0.9769999999999999
- type: precision_at_1000
value: 0.1
- type: precision_at_3
value: 18.54
- type: precision_at_5
value: 13.542000000000002
- type: recall_at_1
value: 30.725
- type: recall_at_10
value: 80.868
- type: recall_at_100
value: 97.653
- type: recall_at_1000
value: 99.57300000000001
- type: recall_at_3
value: 55.619
- type: recall_at_5
value: 67.71000000000001
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-p2p
name: MTEB ArxivClusteringP2P
config: default
split: test
revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d
metrics:
- type: v_measure
value: 44.30960650674069
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-s2s
name: MTEB ArxivClusteringS2S
config: default
split: test
revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53
metrics:
- type: v_measure
value: 38.427074197498996
- task:
type: Reranking
dataset:
type: mteb/askubuntudupquestions-reranking
name: MTEB AskUbuntuDupQuestions
config: default
split: test
revision: 2000358ca161889fa9c082cb41daa8dcfb161a54
metrics:
- type: map
value: 60.28270056031872
- type: mrr
value: 74.38332673789738
- task:
type: STS
dataset:
type: mteb/biosses-sts
name: MTEB BIOSSES
config: default
split: test
revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
metrics:
- type: cos_sim_pearson
value: 84.05942144105269
- type: cos_sim_spearman
value: 82.51212105850809
- type: euclidean_pearson
value: 81.95639829909122
- type: euclidean_spearman
value: 82.3717564144213
- type: manhattan_pearson
value: 81.79273425468256
- type: manhattan_spearman
value: 82.20066817871039
- task:
type: BitextMining
dataset:
type: mteb/bucc-bitext-mining
name: MTEB BUCC (de-en)
config: de-en
split: test
revision: d51519689f32196a32af33b075a01d0e7c51e252
metrics:
- type: accuracy
value: 99.46764091858039
- type: f1
value: 99.37717466945023
- type: precision
value: 99.33194154488518
- type: recall
value: 99.46764091858039
- task:
type: BitextMining
dataset:
type: mteb/bucc-bitext-mining
name: MTEB BUCC (fr-en)
config: fr-en
split: test
revision: d51519689f32196a32af33b075a01d0e7c51e252
metrics:
- type: accuracy
value: 98.29407880255337
- type: f1
value: 98.11248073959938
- type: precision
value: 98.02443319392472
- type: recall
value: 98.2940788025

查看官方来源