Evaluate Your RAG System Performance: Giskard

[@SridharKumarKannam]

If you found this content useful, pleases consider sharing it with others who might benefit. Your support is greatly appreciated :)

[@Jerry-k8b8r]

The Code is missing. Can you please share the code again?

[@pramodpatil1609]

Can we use llama3 for generating test dataset along with as a judge Because i have faced the issue while using llama3

[@SridharKumarKannam]

Yes, we can use any llm. Llama3 worked for me. Whats the issue?

[@pramodpatil1609]

@@SridharKumarKannam issue is Lets Suppose I have provide 10 pdf and create knowledge base and try to create 60 question sets so it wil through error after creating 10 12 question like invalid Json formate by llm like that

[@SridharKumarKannam]

@@pramodpatil1609 use try catch. The output format from llms is esp the smaller models is not always as expected.

[@sangram7153]

can we use our own created test data instead of generating it with Giskard?

[@SridharKumarKannam]

yes you can do that.

[@sangram7153]

please share the code as well

[@SridharKumarKannam]

link added in the description...