✨ feat: 允许用户自行定义 Embedding 模型 #4370

cookieY · 2024-10-14T05:52:19Z

💻 变更类型 | Change Type

🔀 变更说明 | Description of Change

本次变更主要实现了embedding 模型可配置性

通过新增环境变量DEFAULT_FILES_CONFIG实现可自主配置embedding 模型, provider基于现有模型供应商列表

示例:

以 / 为分隔符 openai 为模型提供商, text-embedding-3-small为 embedding 模型

DEFAULT_FILES_CONFIG="embedding_model=zhipu/embedding-3,reranker_model=cohere/rerank-english-v3.0,query_mode=full_text"

📝 补充信息 | Additional Information

目前已支持 zhipu/github/openai/bedrock/ollama模型提供商的embedding 模型

可在agent-runtime下对各模型提供商实现 embeddings 方法从而增加 embedding 模型支持

为后续检索策略预留 rerank 模型以及查询模式变量入口

vercel · 2024-10-14T05:52:22Z

@cookieY is attempting to deploy a commit to the LobeChat Community Team on Vercel.

A member of the Team first needs to authorize it.

lobehubbot · 2024-10-14T05:52:34Z

👍 @cookieY

Thank you for raising your pull request and contributing to our Community
Please make sure you have followed our contributing guidelines. We will review it as soon as possible.
If you encounter any problems, please feel free to connect with us.
非常感谢您提出拉取请求并为我们的社区做出贡献，请确保您已经遵循了我们的贡献指南，我们会尽快审查它。
如果您遇到任何问题，请随时与我们联系。

codecov · 2024-10-14T05:55:18Z

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.21%. Comparing base (780a2a9) to head (f4feb68).
Report is 338 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4370      +/-   ##
==========================================
- Coverage   92.53%   92.21%   -0.32%     
==========================================
  Files         522      571      +49     
  Lines       37335    43419    +6084     
  Branches     3664     2519    -1145     
==========================================
+ Hits        34547    40040    +5493     
- Misses       2788     3379     +591

Flag	Coverage Δ
app	`92.21% <ø> (-0.32%)`	⬇️
server	`97.66% <ø> (+0.39%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

arvinxx

rag_evals 这部分先不要改，目前暂时没有什么人用的，这个 PR 先把最基础的自定义逻辑理顺加上

README.md

.env.example

docs/self-hosting/environment-variables/basic.mdx

src/config/knowledge.ts

src/const/settings/knowledge.ts

src/types/serverConfig.ts

src/libs/agent-runtime/utils/openaiCompatibleFactory/index.ts

arvinxx · 2024-10-28T03:35:41Z

另外 rebase 下 main?

lobehubbot · 2024-10-28T03:35:53Z

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

Also rebase main?

cookieY · 2024-10-28T03:59:58Z

@arvinxx 已根据建议重新更改，解构我会在另提一个 pr

lobehubbot · 2024-10-28T04:00:09Z

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

@arvinxx has changed it according to the suggestion. I will submit another PR to deconstruct it.

arvinxx · 2024-11-05T02:08:09Z

src/config/llm.ts

这几个格式化是不是可以先拆出来 merge 掉

src/server/routers/async/file.ts

src/server/routers/lambda/chunk.ts

BrandonStudio · 2024-11-09T03:34:55Z

模型卡片里面是不是要加embeddingModels？现在只有chatModels
另外models存储是不是要重构？从远程获取模型列表的时候不仅要获取chat模型，还要获取embedding模型（provider的models方法，目前是丢弃所有非chat模型）

lobehubbot · 2024-11-09T03:35:07Z

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

Do I need to add embeddingModels to the model card? Now only chatModels
In addition, does the models storage need to be restructured? When getting the model list from the remote, you need to get not only the chat model, but also the embedding model (provider's models method currently discards all non-chat models)

arvinxx · 2024-11-09T04:09:55Z

模型卡片里面是不是要加embeddingModels？现在只有chatModels
另外models存储是不是要重构？从远程获取模型列表的时候不仅要获取chat模型，还要获取embedding模型（provider的models方法，目前是丢弃所有非chat模型）

嗯，后续是要加的。不过这个还不着急，我们这个月有计划会全量改造一下现在的语言模型那块，会分成供应商配置、语言模型、向量化模型、文生图模型、语音模型等等。

lobehubbot · 2024-11-09T04:10:07Z

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

Do we need to add embeddingModels to the model card? Now only chatModels
In addition, does the models storage need to be restructured? When obtaining the model list remotely, you must obtain not only the chat model, but also the embedding model (provider's models method currently discards all non-chat models)

Well, it will be added later. But don’t worry about it yet. We have plans to completely transform the current language model this month, which will be divided into supplier configuration, language model, vectorization model, Vincentian graph model, speech model, etc.

### [Version 1.35.2](lobehub/lobe-chat@v1.35.1...v1.35.2) <sup>Released on **2024-12-01**</sup> #### 💄 Styles - **misc**: Improve i18n. <br/> <details> <summary><kbd>Improvements and Fixes</kbd></summary> #### Styles * **misc**: Improve i18n, closes [lobehub#4857](lobehub#4857) ([4b7dbc0](lobehub@4b7dbc0)) </details> <div align="right"> [![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top) </div>

lobehubbot · 2024-12-25T03:33:16Z

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

@escapeWu Well, we can move forward.

…tory (lobehub#5116)

### [Version 1.39.2](lobehub/lobe-chat@v1.39.1...v1.39.2) <sup>Released on **2024-12-25**</sup> #### ♻ Code Refactoring - **misc**: Refactor sensenova provider with LobeOpenAICompatibleFactory. <br/> <details> <summary><kbd>Improvements and Fixes</kbd></summary> #### Code refactoring * **misc**: Refactor sensenova provider with LobeOpenAICompatibleFactory, closes [lobehub#5116](lobehub#5116) ([5656f39](lobehub@5656f39)) </details> <div align="right"> [![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top) </div>

* 测试 ppr * fix * fix loading on home * fix style * fix * fix * fix type * try to fix * fix * Delete src/app/(main)/chat/@session/loading.tsx * refactor * fix setting layout * fix loading style * fix import * Update loading.tsx * fix lint

### [Version 1.39.3](lobehub/lobe-chat@v1.39.2...v1.39.3) <sup>Released on **2024-12-25**</sup> #### 💄 Styles - **misc**: Improve loading brand. <br/> <details> <summary><kbd>Improvements and Fixes</kbd></summary> #### Styles * **misc**: Improve loading brand, closes [lobehub#5165](lobehub#5165) ([9684c98](lobehub@9684c98)) </details> <div align="right"> [![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top) </div>

dosubot bot added the size:L This PR changes 100-499 lines, ignoring generated files. label Oct 14, 2024

dosubot bot added the 🌠 Feature Request New feature or request | 特性与建议 label Oct 14, 2024

cookieY mentioned this pull request Oct 14, 2024

✨ feat: 支持用户自行定义 embedding model #4208

Closed

8 tasks

cookieY force-pushed the embedding branch 2 times, most recently from d730ee0 to f60b6f6 Compare October 15, 2024 07:45

This was referenced Oct 18, 2024

[Request] 知识库Ollama Embedding模型支持 #4408

Closed

[Request] 是否可以提供自定義 bedrock embedding model 的範例 #4457

Closed

SpeedupMaster mentioned this pull request Oct 26, 2024

[Bug] 自定义embedding模型为智谱清言，但是依旧无法进行向量化 #4486

Closed

arvinxx reviewed Oct 28, 2024

View reviewed changes

src/libs/agent-runtime/utils/openaiCompatibleFactory/index.ts Outdated Show resolved Hide resolved

dosubot bot added size:XL This PR changes 500-999 lines, ignoring generated files. and removed size:L This PR changes 100-499 lines, ignoring generated files. labels Oct 28, 2024

arvinxx mentioned this pull request Oct 29, 2024

[Request] 自定义嵌入模型 Custom Embedding model #3785

Open

cookieY force-pushed the embedding branch from 72e2b64 to d18e94f Compare October 30, 2024 10:26

dosubot bot added size:L This PR changes 100-499 lines, ignoring generated files. and removed size:XL This PR changes 500-999 lines, ignoring generated files. labels Oct 30, 2024

arvinxx reviewed Nov 5, 2024

View reviewed changes

src/config/llm.ts Outdated

Copy link

Contributor

arvinxx Nov 5, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这几个格式化是不是可以先拆出来 merge 掉

arvinxx reviewed Nov 5, 2024

View reviewed changes

src/server/routers/async/file.ts Outdated Show resolved Hide resolved

src/server/routers/async/file.ts Show resolved Hide resolved

src/server/routers/lambda/chunk.ts Show resolved Hide resolved

hezhijie0327 and others added 21 commits December 25, 2024 11:40

♻️ refactor: refactor sensenova provider with LobeOpenAICompatibleFac…

5656f39

…tory (lobehub#5116)

📝 docs(bot): Auto sync agents & plugin to readme

d73e889

📝 docs(bot): Auto sync agents & plugin to readme

95248dc

✨ feat: 允许用户自行定义 Embedding 模型

ba870f5

✨ feat: customized embedding model support

10dc08e

✨feat: add ollama embedding provider

efda312

🔨 chore: ollama embed multiple texts

03676e2

Not finished changing

125ccf9

🔨 chore: embedding model use DEFAULT_FILES_CONFIG env

39b1d21

add rerank todo

f9aa471

✨ feat: 支持用户自行定义 embedding model 目前支持 zhipu/openai/github/ollama/bedrock

4e03553

📝 docs: Update DEFAULT_FILES_CONFIG

b5db726

♻️refactor: delete unused environment variables

7b751bd

♻️refactor: delete DEFAULT_EMBEDDING_MODEL environment variables

dee1d3f

🔨 chore: add embedding ci file

e0c9b4a

fix: ci file

b3552ad

♻️ refactor: rebase

67e96a7

♻️ refactor: import serverDB in ragEval routers

f4feb68

dosubot bot added size:XXL This PR changes 1000+ lines, ignoring generated files. and removed size:L This PR changes 100-499 lines, ignoring generated files. labels Dec 25, 2024

cookieY closed this Dec 25, 2024

cookieY deleted the embedding branch December 25, 2024 07:48

cookieY restored the embedding branch December 25, 2024 07:48

cookieY deleted the embedding branch December 25, 2024 07:49

cookieY mentioned this pull request Dec 25, 2024

✨ feat: 允许用户自行定义 Embedding 模型 #5177

Open

8 tasks

RS-Imagine mentioned this pull request Dec 29, 2024

[Request] Custom Embedded Model RS-Imagine/lobe-chat#15

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

✨ feat: 允许用户自行定义 Embedding 模型 #4370

✨ feat: 允许用户自行定义 Embedding 模型 #4370

cookieY commented Oct 14, 2024

vercel bot commented Oct 14, 2024

lobehubbot commented Oct 14, 2024

codecov bot commented Oct 14, 2024 •

edited

Loading

arvinxx left a comment

arvinxx commented Oct 28, 2024

lobehubbot commented Oct 28, 2024

cookieY commented Oct 28, 2024

lobehubbot commented Oct 28, 2024

arvinxx Nov 5, 2024

BrandonStudio commented Nov 9, 2024

lobehubbot commented Nov 9, 2024

arvinxx commented Nov 9, 2024

lobehubbot commented Nov 9, 2024

lobehubbot commented Dec 25, 2024

✨ feat: 允许用户自行定义 Embedding 模型 #4370

✨ feat: 允许用户自行定义 Embedding 模型 #4370

Conversation

cookieY commented Oct 14, 2024

💻 变更类型 | Change Type

🔀 变更说明 | Description of Change

📝 补充信息 | Additional Information

vercel bot commented Oct 14, 2024

lobehubbot commented Oct 14, 2024

codecov bot commented Oct 14, 2024 • edited Loading

Codecov Report

arvinxx left a comment

Choose a reason for hiding this comment

arvinxx commented Oct 28, 2024

lobehubbot commented Oct 28, 2024

cookieY commented Oct 28, 2024

lobehubbot commented Oct 28, 2024

arvinxx Nov 5, 2024

Choose a reason for hiding this comment

BrandonStudio commented Nov 9, 2024

lobehubbot commented Nov 9, 2024

arvinxx commented Nov 9, 2024

lobehubbot commented Nov 9, 2024

lobehubbot commented Dec 25, 2024

codecov bot commented Oct 14, 2024 •

edited

Loading