Fix in load_llm.py

[theobgbd]

Description

Fixed an issue where the "proxy" setting was passed to the PublicOpenAPI constructor instead of the "api_base" parameter, disabling the use of on-premise OpenAI-based LLM servers

Related Issues

Issue #1481

Proposed Changes

Passed api_base instead of proxy

Checklist

• I have tested these changes locally.
• I have reviewed the code changes.
• I have updated the documentation (if necessary).
• I have added appropriate unit tests (if applicable).

[HenryZ5734]

Helps a lot👍 With this fix, I can finally run GraphRAG(1.0.0) with local LLM(Qwen-2.5)

[goodpeter-sun]

Thanks for the fix. However, i tried, but it didn't work. Am I missing something?
@theobgbd

[goodpeter-sun]

Helps a lot👍 With this fix, I can finally run GraphRAG(1.0.0) with local LLM(Qwen-2.5)

Are you also using ollama? Could you shed some light? what is your settings.yaml like?

[xfalcox]

Thanks this fixes using it with a LLM deployed via vLLM