How to use OpenAI-style LLM API with this evaluation code

[sikelifei]

Thank you for sharing this project.
The current call_model implementation is tightly coupled to the Qwen-style API (sending prompts and reading from completions). I would like to run the same evaluation pipeline using an OpenAI-compatible LLM API (e.g. OpenAI ChatCompletion / DeepSeek via DashScope compatible mode). how should I adapt the code to use an OpenAI-format LLM endpoint for evaluation?

Thanks!

[ADoublLEN]

Allow us some time and once the development done we will notify you.

If you no hear from me in one week plz knock me again.

[Winlee001]

@ADoublLEN Hi, I'm following up on your previous discussion regarding the OpenAI-compatible API integration—just wondering if there are any updates on the development? I have same need like @sikelifei said,Thanks!

[ADoublLEN]

@ADoublLEN Hi, I'm following up on your previous discussion regarding the OpenAI-compatible API integration—just wondering if there are any updates on the development? I have same need like @sikelifei said,Thanks!

Just done