clinical Intel RAG Model 🏥

Features

• Local LLM: Uses a local Llama-2-7B-chat GGML model for privacy and efficiency.
• Smart Retrieval: Uses SentenceTransformers (all-MiniLM-L6-v2) to find relevant medical diagnoses and lab items.
• Context-Aware Answers: The LLM uses retrieved context to provide accurate answers about medical conditions and lab tests.
• Premium UI: A modern, responsive web interface built with Flask and Vanilla CSS.

Data Sources

• D_ICD_DIAGNOSES.csv: Provides descriptions of thousands of ICD-9 medical codes.
• D_LABITEMS.csv: Provides information on clinical laboratory tests.

Setup

1. Ensure the Llama-2 model is present at: C:\Users\student\Downloads\llama-2-7b-chat.ggmlv3.q2_K.bin
2. Install dependencies:
   pip install -r requirements.txt
3. Run the application:
   python app.py
4. Access the web interface at http://localhost:5000

How it Works

1. Indexing: On startup, the system loads medical knowledge from MIMIC-III CSV files and generates vector embeddings.
2. Retrieval: When a user asks a question, the system finds the most similar medical definitions using cosine similarity.
3. Augmentation: The query and retrieved context are formatted into a prompt for the Llama-2 model.
4. Generation: The LLM generates a helpful, medically-grounded response.