Skip to content

HowToCreateAnAzureOpenAiServiceResourceInAzure.md

Latest commit

 

History

History
34 lines (29 loc) · 2.3 KB

HowToCreateAnAzureOpenAiServiceResourceInAzure.md

File metadata and controls

34 lines (29 loc) · 2.3 KB

How to Create an Azure Open AI Service Resource in Azure

Step 1. How to Create the Resource

  • Go to portal.azure.com and log into you Azure Subscription (If you do not have any, use this guide: https://azure.microsoft.com/en-us/pricing/purchase-options/azure-account)
  • Go to "Hamburger Menu" top Left and choose "Create a Resource"
  • In the "Search service and marketplace" search-box type "Azure openai"
  • Choose "Azure OpenAI", click Create > "Azure OpenAI" (you are take to the creation page)
  • Choose your Subscription and Resource Group (or create a new)
  • Choose your region (We recommend US-East or Swenden Central for most options)
  • Choose a name for the resource (this will be part of your resource URL: https://.openai.azure.com/)
  • Choose Standard S0 (only option) as pricing tier (It is a 100% pay for what you use resource so don't worry about having a dorment resource as there is no fixed price)
  • Click the Next button
  • In Type choose "All Networks"
  • Click Next
  • Leave Tags blank and Click Next
  • Finally Click create (it normally takes less than a minute to complete)
  • Click on "Go to Resource"

Step 2. How to get Endpoint and API Key

  • Go to you Azure OpenAI Resource
  • In the Left Sidebar go to "Resource Management" > "Keys and Endpoint"
  • Note down one of your 2 keys (do not matter which one) and you endpoint (Location/Region is just information and will not be needed in your code) (Should any of your keys be compromized use the "Regenerate Key x" at the top of this page)

Step 3. How to deploy a model

  • Go to you Azure OpenAI Resource and in Overview press "Go to Azure AI Foundery Portal" (or https://ai.azure.com/ for faster access)
  • Once on Azure AI Foundery Portal go to "Deployments" in the left sidebar
  • Press the "+ Deploy Model"
  • Deploy the Model you wish to have available (note down the name you give it as that is needed later in code)
    • We recommend you deploy at least the following models for the Hackathon: "gpt-4o-mini" and "text-embedding-ada-002"
      • Some models are Request Access only. Feel free to request them but it can take up to a month or longer to get the access so do not assume you can try them during the hackathon

(You should now have 3 pieces of info: An Endpoint, a Key and a Model name. If that is the case you are ready to proceed with the code part 🙌)