This GitHub repository demonstrates how to develop Retrieval Augmented Generation(RAG) based chat-bot using Microsoft-Azure OpenAI service. You can refer to code folder of your choice of programming language - Java, Python or Dot Net. Please feel free to reach out to me if you find any bugs or have suggestions for improvement. I strongly recommend to read this documentation before you dive into the code. Happy Azure-AI programming.
First thing first, request for Azure Open-AI service access. Don't worry, initially I had a doubt if my request will be approved? however my request got approved in 24 hours.
Request Access to Azure OpenAI Service
-
In your Azure Portal account search for Azure OpenAI and then select "Azure OpenAI" service.
-
After following step#1 you will see below page. Select "Create Azure OpenAI" button and follow remaining steps.
-
Once you created Azure OpenAI service you go to your Azure OpenAI service resource and then select "Go to Azure OpenAI Studio" link.
-
Following above link you should bring you to this beautiful page.
-
Now select "Deployments" option from left-menu and that should bring to this page.
-
Now create deployments by selecting "Create new Deployment" button. If you notice in below image, I created two deployments. First deployment for text-embeddings using OpenAI "text-embedding-ada-002" model and second deployment for chat/chat-completion using "gpt-35-turbo" model.
Assuming by this time you have your Azure-AI services setup complete and you created model deployments in Azure-OpenAI, lets deep-dive into the code. Please refer to code folder of programming language of your choice above or you can use below link.