Click to use AI to generate Video Script Fails

[mraabhijit]

是否已存在类似问题？

• 我已搜索现有问题

当前行为

I am trying to use the docker image for gpt35 as suggested in the README.md. The app is starting as expected. However, if I try to generate AI script from subject, it fails.

预期行为

The generate AI script should create some sort of/even non-sensical sentence.

重现步骤

Start the docker image for gpt35:

docker run -p 3040:3040 missuo/freegpt35

Once successfully started, modify the config.toml configuration as follows:

• Set llm_provider to openai
• Fill in openai_api_key with any value, for example, '123456'
• Change openai_base_url to http://localhost:3040/v1/
• Set openai_model_name to gpt-3.5-turbo

堆栈追踪/日志

webui  | 2025-02-08 09:40:43 | ERROR | "./app/services/llm.py:371": generate_terms - failed to generate video script: Error: Error code: 401 - {'error': {'message': 'Incorrect API key provided: 123456. You can find your API key at https://platform.openai.com/account/api-keys.', 'type': 'invalid_request_error', 'param': None, 'code': 'invalid_api_key'}}

Python 版本

v11.11

操作系统

Windows 10

MoneyPrinterTurbo 版本

v1.2.2

其他信息

On inspection of the issue with the config.toml, I noticed that the openai_base_url resets to "" even after setting to "http://localhost:3040/v1" which might be the cause of the issue.

This may/may not be due the fact that the docker run of gpt creates a new session every few seconds...

System: Successfully refreshed session ID and token.
System: Successfully refreshed session ID and token.
System: Successfully refreshed session ID and token.
System: Successfully refreshed session ID and token.
System: Successfully refreshed session ID and token.