Skip to content

[Bug] 新版本智能体使用vllm本地部署模型时无法对话启动 #614

@eutopia-ink

Description

@eutopia-ink

Pre-checks

Deployment Method

Docker

Steps to Reproduce

  1. 部署v1.9.2以上的版本
  2. 选择本地vllm部署的大模型创建智能体
  3. 点击新建会话
  4. 提示错误
Image

Expected vs Actual Behavior

对比旧版本创建的agent,socket通信多出了一个onboarding_trigger的报文,

Image

后端处理后,会以system角色消息拼接到llm请求报文的头部,导致llm收到的报文中,有两个system 报文,但vllm不支持,因此返回报错

Image

Logs / Screenshots

底层llm消息拼接类优化,合并相邻system角色消息

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions