/RAG

Primary LanguageJupyter Notebook

README

The data directory contains the data.

RAG contains the Simple RAG and prompt contains code for Multi Agent RAG.

Scrapers contain code for regex and scraping, util files.

Companies: MSFT & AAPL (Microsoft & Apple)

Groq was used for calling LLama-8b. Experiments were also performed using mistral but the API was heavily limited. Langchain & Llamaindex are used for construction of RAG. Gemini was also used but token limited. Llama parser was used to parse the documents.

The app was hosted locally using streamlit.

The Report contains the detailed visualizations and a brief explanation of techniques and results.

Link: https://drive.google.com/file/d/1p8pAMizuhjNfiCwfXWajCXz0SSX1vb-b/view?usp=sharing