Skip to content

Latest commit

 

History

History
66 lines (43 loc) · 2.7 KB

README.md

File metadata and controls

66 lines (43 loc) · 2.7 KB
Raw

RAG with PromptFlow and AISearch

In this example, we will implement a Retrieval Augmented Generation (RAG) application leveraging Phi3 as SLM, AI Search as vectorDB and Prompt Flow as low-code orchestrator.

Features

  • Easy deployment using Docker.
  • Scalable architecture for handling AI workflows.
  • Low code approach using Prompt Flow

Prerequisites

Before you begin, ensure you have met the following requirements:

  • Docker installed on your local machine.
  • An Azure account with permissions to create and manage container resources.
  • An Azure AI Studio and Azure AI Search instances
  • An embedding model to create your index (can be either an Azure OpenAI embedding or an OS model from the catalog)
  • Python 3.8 or later installed on your local machine.
  • An Azure Container Registry (or any registry of your choice)

Installation

  1. Create a new flow on your Azure AI Studio Project using the flow.yaml file.

  2. Deploy a Phi3 Model from your Azure AI model catalog and create the connection to your project. Deploy Phi-3 as a Model as a Service

  3. Create the vector index on Azure AI Search using any document of your choice Create a vector index on Azure AI Search

  4. Deploy the flow on a managed enpoint and use it in the prompt-flow-frontend.py file. Deploy a flow on an online endpoint

  5. Clone the repository:

    git clone [[https://github.com/yourusername/prompt-flow-frontend.git](https://github.com/microsoft/Phi-3CookBook.git)](https://github.com/microsoft/Phi-3CookBook.git)

cd code/07.Lab/RAG with PromptFlow and AISearch

  6. Build the Docker image:

    docker build -t prompt-flow-frontend.py .

  7. Push the Docker image to Azure Container Registry:

    az acr login --name yourregistry

docker tag prompt-flow-frontend.py:latest yourregistry.azurecr.io/prompt-flow-frontend.py:latest

docker push yourregistry.azurecr.io/prompt-flow-frontend.py:latest

Usage

  1. Run the Docker container:

    docker run -p 8501:8501 yourregistry.azurecr.io/prompt-flow-frontend.py:latest

  2. Access the application in your browser at http://localhost:8501.

Contact

Valentina Alto - Linkedin

Full Article: RAG with Phi-3-Medium as a Model as a Service from Azure Model Catalog