常见问题
1、Windows 系统报错configuration failed
出现这个报错时, 可以右键选择“以管理员权限打开”。

2、每次更新版本后建议『重置提供商和模型』
因为可能会有部分模型版本的更新。

3、如果有需要AgentBox执行代码的任务,可以提前下载Python环境(否则对话中遇到此类任务会回复无法执行,请下载Python环境)
下载地址:https://pythonlang.cn/downloads/

4、网络原因报错
- 报错:503 Servic Unavalilable
Ran into this error: Server error:Server error(503 Servic Unavalilable):.Please retryifyouthink this isatransient or recoverable error.
👉网络问题,请切换网络后重试
- 报错:error sending request for url
Ran into this error: Request failed:error sending request for url ( [https://llm-provider.xm-opt.com/v1/chat/completions](https://llm-provider.xm-opt.com/v1/chat/completions) )(failed to connect to IIm-provider.xm-opt.com).
Please retry if you think this is a transient or recoverable error.
👉访问模型服务时,遇到了网络问题,无法连接,可以刷新重试;最常见的是vpn问题,请切换网络后重试。
5、报错:Budget has been exceeded
Ran into this error: Request failed: Bad request (400): Budget has been exceeded! Current cost: 5.033578474869, Max budget: 5.0.
Please retry if you think this is a transient or recoverable error.
👉超过当日限额,请第二天后重试
6、报错:Failed to Load Session
Failed to Load Session
Stream error: Failed to read session:data did not match any variantof untagged enum ResultFormat at line 1 column 222
👉若遇到该问题,请删除当前对话,重新尝试新对话。但之后仍有可能遇到😭,只有不断重试,或更换提问。后续版本将更新
7、报错:Bad request
Ran into this error: Request failed: Bad request (400): litellm.BadRequestError: Vertex_ai_betaException BadRequestError - b'‹\n "error": In "code": 400, \n "message": "Invali‹
JSON payload received. Unexpected token. \n49, 294.57, 288.97, NaN, 301.02, 389.08, \n ^", \n "status": "INVALID_ARGUMENT"\n J\n}\n'.
Please retry if you think this is a transient or recoverable error.
👉可尝试以下任意一种处理方式
- 主动纠正模型推理过程。 发送“数据中有NaN, 请注意做数据清洗或异常处理” 。
- 切换模型重发消息。
- 新开对话, 且切换模型后重发消息。