RAG AI local example

About

RAG AI that uses Phi-3 Mini 4K and All-MiniLM locally

Install

Ollama and Python with virtual environments (venv) 3.12 or latest

ollama run phi3:mini
ollama run all-minilm

pip install -r requirements.txt

Run

uvicorn api_agent_optimized:app --reload

Example

POST endpoint:

http://127.0.0.1:8000/query

json body:

{
    "question": "Qual é a política de home office?"
}