Hi,

HHammad

Hi,
I need to evaluate my RAG pipeline, I used trulens but it is giving me the import error "No module named 'trulens_eval'". I have also tried to reduce the version of trulens but it is not working. Please suggest any other way to evaluate my RAG model, furthermore I am using Gemini as LLM and tried the response evaluation of Llamaindex but it is not providing the satidfactory answer.
The packages version are;
llama-index = 0.9.8
llama-index-core=0.10.30
llama-index-embeddings-gemini =0.1.6
llama-index-llms-gemini =0.1.7
llama-index-multi-modal-llms-gemini= 0.1.5
llama-index-readers-file =0.1.19
llama-index-vector-stores-qdrant=0.2.1
llama-parse = 0.4.1
trulens=1.0.6
trulens-core=1.0.6
trulens-dashboard=1.0.6
trulens-eval = 0.19.1
trulens-feedback=1.0.6

I am open to suggestion. Also I can provide my complete code If required.

3 comments

JJake

Just to clarify your not getting great responses to your queries even with your RAG database? You should be using the Retrieval Evaluation as your looking to judge how relevant the retrieved context is to the query. https://docs.llamaindex.ai/en/stable/examples/evaluation/retrieval/retriever_eval/

HHammad

Thank, let me look into it and if any query arises I will drop a message.

HHammad

Hi Jack,
I am getting the error in the metrics = ["hit_rate", "mrr", "precision", "recall", "ap", "ndcg"] that precision, recall, ap and ndcg are the invalid mertic names. i have upgraded my llamaindex package to 0.11.14. Can you highlight what is the cause of this error?

Add a reply

Find answers from the community

Hi,