search

infiniflow / ragflow

RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding.
https://ragflow.io
Apache License 2.0
22.21k stars 2.18k forks source link

[Question]: it reports the error when i add an tongyi api_key #1748

Open zhanghaiqiangshigezhu opened 3 months ago

zhanghaiqiangshigezhu commented 3 months ago

Describe your problem

it reports the error when i add an tongyi api_key apk_key_error

ARan-forever commented 3 months ago

My situation is not exactly the same as yours. I cannot import the Zhipu API. And I have successfully deployed it on a laptop, but I just can't deploy it on this new computer.

wrong

It shows that the token has expired, but my Zhipu official website api interface clearly has not expired. I don’t understand why this problem occurs.

godeamon commented 3 months ago

我遇到同样的问题了，不清楚为什么默认是qwen-max-1201这个模型，及时我改成了其他模型。而且添加APIKey一直报错请求太频繁。实际上看日志只有一个请求。

zryf2000 commented 3 months ago

qwen-max-1201 该模型为qwen-max的2023年12月1号的历史快照固定版本，该版本的维护时间已经到期，将在4月22日下线，请及时迁移到更新版本模型。（4月8日开始模型限流也会逐步调低直至下线）。

godeamon commented 3 months ago

qwen-max-1201 该模型为qwen-max的2023年12月1号的历史快照固定版本，该版本的维护时间已经到期，将在4月22日下线，请及时迁移到更新版本模型。（4月8日开始模型限流也会逐步调低直至下线）。

感谢回复，问题是即使我在页面选择了其他模型，但是添加 APIKey 时还是会有这个模型的错误。我怀疑是默认就选择了这个模型。请问如何修改默认模型？

zryf2000 commented 3 months ago

bug

KevinHuSh commented 3 months ago

I guess your account credit has limitation of call frequency.

Phillee80 commented 3 months ago

I guess your account credit has limitation of call frequency.

when add LLM model it only asks for API key, there is no way to choose which Qwen model i'd like to choose. Actually i dont want use Qwen-max-1201, it got stuck

xxz196801 commented 3 months ago

同上：同样的问题 提示 : 102 Fail to access model(qwen-max-1201) using this api key.ERROR: Requests rate limit exceeded, please try again later.

cwmine commented 3 months ago

同样的问题

cwmine commented 3 months ago

Is any temporary solution?

20070951 commented 3 months ago

Why do you guys just paste your api code here？

jinguoyang commented 1 month ago

RAGFlow在替换/新增api-key的时候，会尝试对内置模型进行轮询访问，以确认key有效。由于通义已经不再提供qwen-max-1201模型服务，因此程序无法得到正确response，认为其api-key有问题，导致新增api-key失败。

内置模型在mysql数据库中进行管理。因此可将数据库llm表中的qwen-max-1201模型强制删除。

数据库的登录信息在docker/service_conf.yaml中。

In RAGFlow, when replacing or adding a new API key, the system performs a polling check on the built-in models to validate the key. However, since the qwen-max-1201 model service is no longer provided by Tongyi, the program is unable to receive a valid response. This leads the system to mistakenly flag the API key as invalid, resulting in a failure to add the new key.

The built-in models are managed through the MySQL database, where the relevant configurations are stored. To resolve the issue, you can manually remove the qwen-max-1201 model from the llm table in the database.

Database login credentials can be found in the docker/service_conf.yaml file.