multilingual e5 base

提供商intfloat
分类sentence-similarity
许可证mit
下载量1.8K
星标0

简介

multilingual-e5-base 是一款高性能的多语言文本向量化模型，由 intfloat 发布。它将文本转化为高维向量，能够精准捕捉语义相似度，而非简单的关键词匹配。对于中国开发者而言，该模型最大的价值在于其强大的跨语言能力，能有效处理中英混合文本，是构建 RAG（检索增强生成）系统、语义搜索或智能知识库的理想底层工具。相比于调用昂贵的闭源 API，它体积适中，部署门槛低，且在多语言语义对齐上表现稳健，能显著提升本地知识库的检索召回率。

核心亮点

  • 支持多语言语义对齐，中英互检效果出色
  • RAG 架构的核心组件，大幅提升检索精准度
  • MIT 协议开源，可低成本私有化部署
  • 语义相似度计算快，适用大规模文档聚类

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-base")
tokenizer = AutoTokenizer.from_pretrained("intfloat/multilingual-e5-base")

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-base

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-base 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-base')

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-base

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-base

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-base')
tokenizer = AutoTokenizer.from_pretrained('intfloat/multilingual-e5-base')

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-base

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-base 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-base')

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-base.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-base.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-base')

完整文档

来源: HuggingFace

---
tags:

  • mteb

  • Sentence Transformers

  • sentence-similarity

  • sentence-transformers

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

results:
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (en)
config: en
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 78.97014925373135
- type: ap
value: 43.69351129103008
- type: f1
value: 73.38075030070492
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (de)
config: de
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 71.7237687366167
- type: ap
value: 82.22089859962671
- type: f1
value: 69.95532758884401
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (en-ext)
config: en-ext
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 79.65517241379312
- type: ap
value: 28.507918657094738
- type: f1
value: 66.84516013726119
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (ja)
config: ja
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 73.32976445396146
- type: ap
value: 20.720481637566014
- type: f1
value: 59.78002763416003
- task:
type: Classification
dataset:
type: mteb/amazon_polarity
name: MTEB AmazonPolarityClassification
config: default
split: test
revision: e2d317d38cd51312af73b3d32a06d1a08b442046
metrics:
- type: accuracy
value: 90.63775
- type: ap
value: 87.22277903861716
- type: f1
value: 90.60378636386807
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (en)
config: en
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 44.546
- type: f1
value: 44.05666638370923
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (de)
config: de
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 41.828
- type: f1
value: 41.2710255644252
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (es)
config: es
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 40.534
- type: f1
value: 39.820743174270326
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (fr)
config: fr
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 39.684
- type: f1
value: 39.11052682815307
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (ja)
config: ja
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 37.436
- type: f1
value: 37.07082931930871
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (zh)
config: zh
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 37.226000000000006
- type: f1
value: 36.65372077739185
- task:
type: Retrieval
dataset:
type: arguana
name: MTEB ArguAna
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 22.831000000000003
- type: map_at_10
value: 36.42
- type: map_at_100
value: 37.699
- type: map_at_1000
value: 37.724000000000004
- type: map_at_3
value: 32.207
- type: map_at_5
value: 34.312
- type: mrr_at_1
value: 23.257
- type: mrr_at_10
value: 36.574
- type: mrr_at_100
value: 37.854
- type: mrr_at_1000
value: 37.878
- type: mrr_at_3
value: 32.385000000000005
- type: mrr_at_5
value: 34.48
- type: ndcg_at_1
value: 22.831000000000003
- type: ndcg_at_10
value: 44.230000000000004
- type: ndcg_at_100
value: 49.974000000000004
- type: ndcg_at_1000
value: 50.522999999999996
- type: ndcg_at_3
value: 35.363
- type: ndcg_at_5
value: 39.164
- type: precision_at_1
value: 22.831000000000003
- type: precision_at_10
value: 6.935
- type: precision_at_100
value: 0.9520000000000001
- type: precision_at_1000
value: 0.099
- type: precision_at_3
value: 14.841
- type: precision_at_5
value: 10.754
- type: recall_at_1
value: 22.831000000000003
- type: recall_at_10
value: 69.346
- type: recall_at_100
value: 95.235
- type: recall_at_1000
value: 99.36
- type: recall_at_3
value: 44.523
- type: recall_at_5
value: 53.769999999999996
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-p2p
name: MTEB ArxivClusteringP2P
config: default
split: test
revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d
metrics:
- type: v_measure
value: 40.27789869854063
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-s2s
name: MTEB ArxivClusteringS2S
config: default
split: test
revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53
metrics:
- type: v_measure
value: 35.41979463347428
- task:
type: Reranking
dataset:
type: mteb/askubuntudupquestions-reranking
name: MTEB AskUbuntuDupQuestions
config: default
split: test
revision: 2000358ca161889fa9c082cb41daa8dcfb161a54
metrics:
- type: map
value: 58.22752045109304
- type: mrr
value: 71.51112430198303
- task:
type: STS
dataset:
type: mteb/biosses-sts
name: MTEB BIOSSES
config: default
split: test
revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
metrics:
- type: cos_sim_pearson
value: 84.71147646622866
- type: cos_sim_spearman
value: 85.059167046486
- type: euclidean_pearson
value: 75.88421613600647
- type: euclidean_spearman
value: 75.12821787150585
- type: manhattan_pearson
value: 75.22005646957604
- type: manhattan_spearman
value: 74.42880434453272
- 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: 99.23799582463465
- type: f1
value: 99.12665274878218
- type: precision
value: 99.07098121085595
- type: recall
value: 99.23799582463465
- 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: 97.88685890380806
- type: f1
value: 97.59336708489249
- type: precision
value: 97.44662117543473
- type: recall
valu

查看官方来源