nomic embed text v1.5

提供商nomic-ai
分类sentence-similarity
许可证apache-2.0
下载量20.9K
星标10

简介

nomic embed text v1.5 是一款高性能的开源文本向量化模型，主要用于将文本转换为高维向量，以便计算机理解语义相似度。它在中文等多语言支持和长文本处理上表现稳健，是构建 RAG（检索增强生成）系统的核心组件。相比于闭源的 OpenAI embedding 接口，它不仅能本地部署以保证数据隐私，且在很多基准测试中性能相当。对于开发者来说，它能无缝集成到 LangChain 或 LlamaIndex 等主流框架中，上手门槛极低，是替代商业 API、降低推理成本的理想选择。

核心亮点

  • 支持长文本输入，显著提升 RAG 检索精度
  • 开源 Apache-2.0 协议，支持私有化部署
  • 语义匹配能力强，完美替代商业 Embedding API
  • 兼容主流 AI 框架，快速集成至知识库项目

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("nomic-ai/nomic-embed-text-v1.5")
tokenizer = AutoTokenizer.from_pretrained("nomic-ai/nomic-embed-text-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 nomic-ai/nomic-embed-text-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 nomic-ai/nomic-embed-text-v1.5 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('nomic-ai/nomic-embed-text-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/nomic-ai/nomic-embed-text-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/nomic-ai/nomic-embed-text-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('nomic-ai/nomic-embed-text-v1.5')
tokenizer = AutoTokenizer.from_pretrained('nomic-ai/nomic-embed-text-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 nomic-ai/nomic-embed-text-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 nomic-ai/nomic-embed-text-v1.5 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('nomic-ai/nomic-embed-text-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/nomic-ai/nomic-embed-text-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/nomic-ai/nomic-embed-text-v1.5.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', 'nomic-ai/nomic-embed-text-v1.5')

完整文档

来源: HuggingFace

---
library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:

  • feature-extraction

  • sentence-similarity

  • mteb

  • transformers

  • transformers.js

model-index:
  • name: epoch_0_model

results:
- task:
type: Classification
dataset:
type: mteb/amazon_counterfactual
name: MTEB AmazonCounterfactualClassification (en)
config: en
split: test
revision: e8379541af4e31359cca9fbcf4b00f2671dba205
metrics:
- type: accuracy
value: 75.20895522388058
- type: ap
value: 38.57605549557802
- type: f1
value: 69.35586565857854
- task:
type: Classification
dataset:
type: mteb/amazon_polarity
name: MTEB AmazonPolarityClassification
config: default
split: test
revision: e2d317d38cd51312af73b3d32a06d1a08b442046
metrics:
- type: accuracy
value: 91.8144
- type: ap
value: 88.65222882032363
- type: f1
value: 91.80426301643274
- task:
type: Classification
dataset:
type: mteb/amazon_reviews_multi
name: MTEB AmazonReviewsClassification (en)
config: en
split: test
revision: 1399c76144fd37290681b995c656ef9b2e06e26d
metrics:
- type: accuracy
value: 47.162000000000006
- type: f1
value: 46.59329642263158
- task:
type: Retrieval
dataset:
type: arguana
name: MTEB ArguAna
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 24.253
- type: map_at_10
value: 38.962
- type: map_at_100
value: 40.081
- type: map_at_1000
value: 40.089000000000006
- type: map_at_3
value: 33.499
- type: map_at_5
value: 36.351
- type: mrr_at_1
value: 24.609
- type: mrr_at_10
value: 39.099000000000004
- type: mrr_at_100
value: 40.211000000000006
- type: mrr_at_1000
value: 40.219
- type: mrr_at_3
value: 33.677
- type: mrr_at_5
value: 36.469
- type: ndcg_at_1
value: 24.253
- type: ndcg_at_10
value: 48.010999999999996
- type: ndcg_at_100
value: 52.756
- type: ndcg_at_1000
value: 52.964999999999996
- type: ndcg_at_3
value: 36.564
- type: ndcg_at_5
value: 41.711999999999996
- type: precision_at_1
value: 24.253
- type: precision_at_10
value: 7.738
- type: precision_at_100
value: 0.98
- type: precision_at_1000
value: 0.1
- type: precision_at_3
value: 15.149000000000001
- type: precision_at_5
value: 11.593
- type: recall_at_1
value: 24.253
- type: recall_at_10
value: 77.383
- type: recall_at_100
value: 98.009
- type: recall_at_1000
value: 99.644
- type: recall_at_3
value: 45.448
- type: recall_at_5
value: 57.965999999999994
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-p2p
name: MTEB ArxivClusteringP2P
config: default
split: test
revision: a122ad7f3f0291bf49cc6f4d32aa80929df69d5d
metrics:
- type: v_measure
value: 45.69069567851087
- task:
type: Clustering
dataset:
type: mteb/arxiv-clustering-s2s
name: MTEB ArxivClusteringS2S
config: default
split: test
revision: f910caf1a6075f7329cdf8c1a6135696f37dbd53
metrics:
- type: v_measure
value: 36.35185490976283
- task:
type: Reranking
dataset:
type: mteb/askubuntudupquestions-reranking
name: MTEB AskUbuntuDupQuestions
config: default
split: test
revision: 2000358ca161889fa9c082cb41daa8dcfb161a54
metrics:
- type: map
value: 61.71274951450321
- type: mrr
value: 76.06032625423207
- task:
type: STS
dataset:
type: mteb/biosses-sts
name: MTEB BIOSSES
config: default
split: test
revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
metrics:
- type: cos_sim_pearson
value: 86.73980520022269
- type: cos_sim_spearman
value: 84.24649792685918
- type: euclidean_pearson
value: 85.85197641158186
- type: euclidean_spearman
value: 84.24649792685918
- type: manhattan_pearson
value: 86.26809552711346
- type: manhattan_spearman
value: 84.56397504030865
- task:
type: Classification
dataset:
type: mteb/banking77
name: MTEB Banking77Classification
config: default
split: test
revision: 0fd18e25b25c072e09e0d92ab615fda904d66300
metrics:
- type: accuracy
value: 84.25324675324674
- type: f1
value: 84.17872280892557
- task:
type: Clustering
dataset:
type: mteb/biorxiv-clustering-p2p
name: MTEB BiorxivClusteringP2P
config: default
split: test
revision: 65b79d1d13f80053f67aca9498d9402c2d9f1f40
metrics:
- type: v_measure
value: 38.770253446400886
- task:
type: Clustering
dataset:
type: mteb/biorxiv-clustering-s2s
name: MTEB BiorxivClusteringS2S
config: default
split: test
revision: 258694dd0231531bc1fd9de6ceb52a0853c6d908
metrics:
- type: v_measure
value: 32.94307095497281
- task:
type: Retrieval
dataset:
type: BeIR/cqadupstack
name: MTEB CQADupstackAndroidRetrieval
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 32.164
- type: map_at_10
value: 42.641
- type: map_at_100
value: 43.947
- type: map_at_1000
value: 44.074999999999996
- type: map_at_3
value: 39.592
- type: map_at_5
value: 41.204
- type: mrr_at_1
value: 39.628
- type: mrr_at_10
value: 48.625
- type: mrr_at_100
value: 49.368
- type: mrr_at_1000
value: 49.413000000000004
- type: mrr_at_3
value: 46.400000000000006
- type: mrr_at_5
value: 47.68
- type: ndcg_at_1
value: 39.628
- type: ndcg_at_10
value: 48.564
- type: ndcg_at_100
value: 53.507000000000005
- type: ndcg_at_1000
value: 55.635999999999996
- type: ndcg_at_3
value: 44.471
- type: ndcg_at_5
value: 46.137
- type: precision_at_1
value: 39.628
- type: precision_at_10
value: 8.856
- type: precision_at_100
value: 1.429
- type: precision_at_1000
value: 0.191
- type: precision_at_3
value: 21.268
- type: precision_at_5
value: 14.649000000000001
- type: recall_at_1
value: 32.164
- type: recall_at_10
value: 59.609
- type: recall_at_100
value: 80.521
- type: recall_at_1000
value: 94.245
- type: recall_at_3
value: 46.521
- type: recall_at_5
value: 52.083999999999996
- task:
type: Retrieval
dataset:
type: BeIR/cqadupstack
name: MTEB CQADupstackEnglishRetrieval
config: default
split: test
revision: None
metrics:
- type: map_at_1
value: 31.526
- type: map_at_10
value: 41.581
- type: map_at_100
value: 42.815999999999995
- type: map_at_1000
value: 42.936
- type: map_at_3
value: 38.605000000000004
- type: map_at_5
value: 40.351
- type: mrr_at_1
value: 39.489999999999995
- type: mrr_at_10
value: 47.829
- type: mrr_at_100
value: 48.512
- type: mrr_at_1000
value: 48.552
- type: mrr_at_3
value: 45.754
- type: mrr_at_5
value: 46.986
- type: ndcg_at_1
value: 39.489999999999995
- type: ndcg_at_10
value: 47.269
- type: ndcg_at_100
value: 51.564
- type: ndcg_at_1000
value: 53.53099999999999
- type: ndcg_at_3
value: 43.301
- type: ndcg_at_5
value: 45.239000000000004
- type: precision_at_1
value: 39.489999999999995
- type: precision_at_10
value: 8.93
- type:

查看官方来源