update README

README.md

@@ -1,2 +1,9 @@
+Before using the TAIDE llm model (in the RAG_strategy_Taide.py), run below first.
+
 run `chmod +x run.sh`
-run `./run.sh`
+
+run `./run.sh`
+
+It will download and build the model.
+
+The `faiss_index` folder requires the Taide model.

faiss_index/README_faiss.md

@@ -1,9 +1,13 @@
-Comparing Faiss with the original RAG. Using the TAIDE LLM model. Using created embeddings in the Supabase db.
+Comparing Faiss with the original RAG. Both using the TAIDE LLM model. Both using created embeddings from the Supabase db.
 
-Did not apply multi_query on Faiss (it will generate very similar answers from the original RAG). Applied multi_query on the original RAG only.
+Did not apply multi_query on Faiss (as it generates very similar answers from the original RAG). Applied multi_query on the original RAG only.
 
 Evaluate on: 1. speed 2. RAGAS score
-Some of the result is in the output.txt
 
+Some of the tested result is in the output.txt
+
+Findings:
+- Response speed: It's not worth using Faiss to speed up llm on small dataset. Its index method is designed for large dataset. 
+- Response quality: If relevant documents can be found, the original RAG which spplied the multi query method can generate more detailed answers. Faiss' answers are shorter.  
 
 Faiss: https://engineering.fb.com/2017/03/29/data-infrastructure/faiss-a-library-for-efficient-similarity-search/