Good morning,
I'm currently working on the RAG Project and I wanted to ask for some clarifications, particularly on things I should do to carry out a good project.
I have currently worked/ I'm working on the following points:
- Selection and testing of different embedding models
- Selection and testing of different LLMs
- Selection and testing of various parameters (chunk sizes, temperature, etc.)
- See how the answers (at the same questions) vary with different parameters setÂ
- Considerations on the tests done
However, I wondered if I was forgetting something or if there was something else to work on that I had missed.
I ask this because I would not like to deliver a notebook that is too similar to the example given (the Barbienheimer one).
Maybe this discussion can be helpful to exchange ideas on how to diversify projects and intervene on the code.
Thank you very much and have you all a good day!