Small RAG assistant for the Rules of warhammer 40K
Install the requirements :
pip install -r requirements.txt
-
Download an embedding model like allMiniLM-L6-v2 from huggingface
-
Donwload a llm model supported by llamaCPP (I used zephyr-7b-beta.Q3_K_S)
-
Indicate the locations of the models in the
streamlit_app.pyscript:EMBED_MODEL = <PATH TO YOUR EMBEDING MODEL LLM_MODEL = <PATH TO YOUR LLM MODEL>
run the app with the command
streamlit run ./streamlit_app.py
If you run the program locally with limited resources. It can take a few minute to get a result. Most of the results are of good quality even though it can skip some details.
