bge base en v1.5
简介
bge-base-en-v1.5 是由北京智源人工智能研究院（BAAI）推出的轻量级英文向量模型。它专注于将文本转化为高质量的向量表示（Embedding），是构建 RAG（检索增强生成）系统的核心组件。相比于大型 LLM，它的运行开销极低，上手门槛几乎为零，主要用于语义搜索、文档聚类和知识库检索。在实际开发中，它通常与 Milvus 或 Pinecone 等向量数据库配合使用，作为“检索端”来精准筛选相关上下文，从而提升大模型回答的准确性。
核心亮点
- 由智源研究院开发，英文语义检索能力强
- 轻量化部署，推理速度快且资源占用低
- RAG 架构首选，显著提升知识库检索精度
- 采用 MIT 协议，对商业化应用非常友好
使用方法
安装依赖
# 安装 Hugging Face transformers
pip install transformers torchSDK 使用
# 使用 transformers 加载模型
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("BAAI/bge-base-en-v1.5")
tokenizer = AutoTokenizer.from_pretrained("BAAI/bge-base-en-v1.5")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 BAAI/bge-base-en-v1.5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 BAAI/bge-base-en-v1.5 config.json --local-dir ./dirSDK 下载
SDK 下载
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('BAAI/bge-base-en-v1.5')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/BAAI/bge-base-en-v1.5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/BAAI/bge-base-en-v1.5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('BAAI/bge-base-en-v1.5')
tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-base-en-v1.5')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 BAAI/bge-base-en-v1.5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 BAAI/bge-base-en-v1.5 README.md --local_dir ./dirSDK 下载
SDK 下载
# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('BAAI/bge-base-en-v1.5')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/BAAI/bge-base-en-v1.5.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/BAAI/bge-base-en-v1.5.gitModelScope 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', 'BAAI/bge-base-en-v1.5')完整文档
来源: HuggingFace
---
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
- mteb
model-index:
- name: bge-base-en-v1.5
results:
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (en)
config: en
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 76.14925373134328
- type: ap
value: 39.32336517995478
- type: f1
value: 70.16902252611425
- task:
type: Classification
dataset:
type: mteb/amazon_polarity
name: MTEB AmazonPolarityClassification
config: default
split: test
revision: e2d317d38cd51312af73b3d32a06d1a08b442046
metrics:
- type: accuracy
value: 93.386825
- type: ap
value: 90.21276917991995
- type: f1
value: 93.37741030006174
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (en)
config: en
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 48.846000000000004
- type: f1
value: 48.14646269778261
- task:
type: Retrieval
dataset:
type: arguana
name: MTEB ArguAna
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 40.754000000000005
- type: map_at_10
value: 55.761
- type: map_at_100
value: 56.330999999999996
- type: map_at_1000
value: 56.333999999999996
- type: map_at_3
value: 51.92
- type: map_at_5
value: 54.010999999999996
- type: mrr_at_1
value: 41.181
- type: mrr_at_10
value: 55.967999999999996
- type: mrr_at_100
value: 56.538
- type: mrr_at_1000
value: 56.542
- type: mrr_at_3
value: 51.980000000000004
- type: mrr_at_5
value: 54.208999999999996
- type: ndcg_at_1
value: 40.754000000000005
- type: ndcg_at_10
value: 63.605000000000004
- type: ndcg_at_100
value: 66.05199999999999
- type: ndcg_at_1000
value: 66.12
- type: ndcg_at_3
value: 55.708
- type: ndcg_at_5
value: 59.452000000000005
- type: precision_at_1
value: 40.754000000000005
- type: precision_at_10
value: 8.841000000000001
- type: precision_at_100
value: 0.991
- type: precision_at_1000
value: 0.1
- type: precision_at_3
value: 22.238
- type: precision_at_5
value: 15.149000000000001
- type: recall_at_1
value: 40.754000000000005
- type: recall_at_10
value: 88.407
- type: recall_at_100
value: 99.14699999999999
- type: recall_at_1000
value: 99.644
- type: recall_at_3
value: 66.714
- type: recall_at_5
value: 75.747
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-p2p
name: MTEB ArxivClusteringP2P
config: default
split: test
revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d
metrics:
- type: v_measure
value: 48.74884539679369
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-s2s
name: MTEB ArxivClusteringS2S
config: default
split: test
revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53
metrics:
- type: v_measure
value: 42.8075893810716
- task:
type: Reranking
dataset:
type: mteb/askubuntudupquestions-reranking
name: MTEB AskUbuntuDupQuestions
config: default
split: test
revision: 2000358ca161889fa9c082cb41daa8dcfb161a54
metrics:
- type: map
value: 62.128470519187736
- type: mrr
value: 74.28065778481289
- task:
type: STS
dataset:
type: mteb/biosses-sts
name: MTEB BIOSSES
config: default
split: test
revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
metrics:
- type: cos_sim_pearson
value: 89.24629081484655
- type: cos_sim_spearman
value: 86.93752309911496
- type: euclidean_pearson
value: 87.58589628573816
- type: euclidean_spearman
value: 88.05622328825284
- type: manhattan_pearson
value: 87.5594959805773
- type: manhattan_spearman
value: 88.19658793233961
- task:
type: Classification
dataset:
type: mteb/banking77
name: MTEB Banking77Classification
config: default
split: test
revision: 0fd18e25b25c072e09e0d92ab615fda904d66300
metrics:
- type: accuracy
value: 86.9512987012987
- type: f1
value: 86.92515357973708
- task:
type: Clustering
dataset:
type: mteb/biorxiv-clustering-p2p
name: MTEB BiorxivClusteringP2P
config: default
split: test
revision: 65b79d1d13f80053f67aca9498d9402c2d9f1f40
metrics:
- type: v_measure
value: 39.10263762928872
- task:
type: Clustering
dataset:
type: mteb/biorxiv-clustering-s2s
name: MTEB BiorxivClusteringS2S
config: default
split: test
revision: 258694dd0231531bc1fd9de6ceb52a0853c6d908
metrics:
- type: v_measure
value: 36.69711517426737
- task:
type: Retrieval
dataset:
type: BeIR/cqadupstack
name: MTEB CQADupstackAndroidRetrieval
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 32.327
- type: map_at_10
value: 44.099
- type: map_at_100
value: 45.525
- type: map_at_1000
value: 45.641999999999996
- type: map_at_3
value: 40.47
- type: map_at_5
value: 42.36
- type: mrr_at_1
value: 39.199
- type: mrr_at_10
value: 49.651
- type: mrr_at_100
value: 50.29
- type: mrr_at_1000
value: 50.329
- type: mrr_at_3
value: 46.924
- type: mrr_at_5
value: 48.548
- type: ndcg_at_1
value: 39.199
- type: ndcg_at_10
value: 50.773
- type: ndcg_at_100
value: 55.67999999999999
- type: ndcg_at_1000
value: 57.495
- type: ndcg_at_3
value: 45.513999999999996
- type: ndcg_at_5
value: 47.703
- type: precision_at_1
value: 39.199
- type: precision_at_10
value: 9.914000000000001
- type: precision_at_100
value: 1.5310000000000001
- type: precision_at_1000
value: 0.198
- type: precision_at_3
value: 21.984
- type: precision_at_5
value: 15.737000000000002
- type: recall_at_1
value: 32.327
- type: recall_at_10
value: 63.743
- type: recall_at_100
value: 84.538
- type: recall_at_1000
value: 96.089
- type: recall_at_3
value: 48.065000000000005
- type: recall_at_5
value: 54.519
- task:
type: Retrieval
dataset:
type: BeIR/cqadupstack
name: MTEB CQADupstackEnglishRetrieval
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 32.671
- type: map_at_10
value: 42.954
- type: map_at_100
value: 44.151
- type: map_at_1000
value: 44.287
- type: map_at_3
value: 39.912
- type: map_at_5
value: 41.798
- type: mrr_at_1
value: 41.465
- type: mrr_at_10
value: 49.351
- type: mrr_at_100
value: 49.980000000000004
- type: mrr_at_1000
value: 50.016000000000005
- type: mrr_at_3
value: 47.144000000000005
- type: mrr_at_5
value: 48.592999999999996
- type: ndcg_at_1
value: 41.465
- type: ndcg_at_10
value: 48.565999999999995
- type: ndcg_at_100
value: 52.76499999999999
- type: ndcg_at_1000
value: 54.749
- type: ndcg_at_3
value: 44.57
- type: ndcg_at_5
value: 46.759
- type: precision_at_1
value: 41.465
- type: precision_at_10
value:
