How can I change model from "gpt-4o-mini" to any ollama model in Evaluation ?

[vron8632]

I don't have a valid OPENAI_API_KEY, so I want to use local ollama model such as qwen3 to reproduce "winrate_comparison" and "quantitative_comparison"

[Re-bin]

Hi 👋!

Thanks for your interest in VideoRAG! If you want to reproduce our results, it’s best to use the same model we used (i.e., gpt-4o-mini).

But if you just want to try different RAG models to see which one gives better answers for you, go ahead and use whichever model you like.

Hope this helps:)

Best regards,
Xubin