Skip to content

[Bug] 【向量模型】通过openAI配置本地vllm部署的Qwen3-Embedding-8B向量化模型，文件向量化失败。 #5295

Closed
Closed
[Bug] 【向量模型】通过openAI配置本地vllm部署的Qwen3-Embedding-8B向量化模型，文件向量化失败。#5295
Assignees
zyyfit

Description

@2lipan
2lipan
opened on May 25, 2026

Contact Information

No response

MaxKB Version

v2.9.2

Problem Description

【向量模型】通过本地vllm部署的Qwen3-Embedding-8B向量化模型，在页面通过openAI配置进去，然后创建一个知识库，使用这个向量化模型，会向量化失败。模型日志里一直显示 "POST /v1/embeddings HTTP/1.1" 400 Bad Request 错误

Steps to Reproduce

【向量模型】通过本地vllm部署的Qwen3-Embedding-8B向量化模型，在页面通过openAI配置进去，然后创建一个知识库，使用这个向量化模型，会向量化失败。模型日志里一直显示 "POST /v1/embeddings HTTP/1.1" 400 Bad Request 错误

Image Image Image

在服务器上通过命令行执行向量化是成功的

Image Image

The expected correct result

No response

Related log output



Additional Information


No response
Activity
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Metadata
Metadata
Assignees
Labels
No labels
No labels
Type
No type
Projects
No projects
    Milestone
    No milestone
    Relationships
    None yet
    Development
    No branches or pull requests
    Issue actions