eosphoros-ai / DB-GPT

AI Native Data App Development framework with AWEL(Agentic Workflow Expression Language) and Agents
http://docs.dbgpt.cn
MIT License
13.24k stars 1.76k forks source link

[Feature][ChatKnowledge] The parameters of the reranker model are configurable.[reranker 模型参数可配置] #1822

Open SuperKevinZhang opened 1 month ago

SuperKevinZhang commented 1 month ago

Search before asking

Description

Current Situation: Typically, different knowledge bases have different requirements for the recall parameters. Currently, the configuration interface for the knowledge base only includes the recall parameters for the Embedding model and lacks the parameter settings for the reranker model.

Expectation: Add reranker model parameter settings, such as model name and topK parameter, to the knowledge base configuration interface.

现状:通常情况下,不同知识库对召回的参数有不同的要求,目前知识库配置画面只有Embedding的召回参数设置,缺少reranker模型参数设置; 期望:在知识库画面,增加reranker模型的参数设置,如模型名字,topK参数等。

Use case

No response

Related issues

No response

Feature Priority

Medium

Are you willing to submit PR?

Aries-ckt commented 4 weeks ago

@SuperKevinZhang thanks for your suggestion, we will integrate rerank param soon.