where all am i wrong??

Question

Azure Ai Services(used)

Ai search

Creation : service/resource is created under the Azure -> Azure Ai Services -> AI Search. 

For searching through the DataBase, Vector database. An embedding model needs to be deployed in order to use this.

Open-ai-ada model was used.

This service also provides a Database to store data.

Configurations in the DB : search algorithm used, hybrid , semantic or vector 

Azure Ai Services(same name as parent)

Creation : service/resource is created under the Azure -> Azure Ai Services -> Azure ai se services(child). Let's call this created resource “AAS”

Created along the way : text, audio, visual model service. 

And an AI foundry is created for you under this same resource. Where you can launch this “AAS”. To access this you need Azure Ai Services -> Azure Ai Services -> Resource name(AAS) -> go to AI foundry portal()

But this foundry will only give you access to 62 models and all of them chat completion models.

If so, I Don't know how they’ll provide the speech, vision or any other service they mentioned in the bottom page of “AAS”.

But the funny thing is if we create a resource in AI foundry service separately(let's call it “AIDRY”)it will fall under this Azure AI service(child), and will be available alongside AAS.

And if you try to open “AIDRY” in AI foundry from Azure AI service(child) it will have a slightly different button then AAS to open the foundry. And will offer all the 1800 models of all kinds.

Azure open AI

Creation : service/resource is created under the Azure -> Azure Ai Services -> Azure OpenAi

  Just creates a Open Service resource, then allows you to pick any model for any functionality 
  

1. Azure AI foundry
   • Creation : service/resource is created under the Azure -> Azure AI Foundry. Let's call this created resource “AAF”.
   • This will take you directly to the foundry, where you can again create an Azure AI service(child).
   • But the difference is you will have access to all the 1800+ models when you open the foundry.
   • And this will automatically have the resources like speech and vision which are created along the way just like an Azure Ai service would do.

Azure Ai Services(used)

Ai search

Creation : service/resource is created under the Azure -> Azure Ai Services -> AI Search. 

For searching through the DataBase, Vector database. An embedding model needs to be deployed in order to use this.

Open-ai-ada model was used.

This service also provides a Database to store data.

Configurations in the DB : search algorithm used, hybrid , semantic or vector 

Azure Ai Services(same name as parent)

Creation : service/resource is created under the Azure -> Azure Ai Services -> Azure ai se services(child). Let's call this created resource “AAS”

Created along the way : text, audio, visual model service. 

And an AI foundry is created for you under this same resource. Where you can launch this “AAS”. To access this you need Azure Ai Services -> Azure Ai Services -> Resource name(AAS) -> go to AI foundry portal()

But this foundry will only give you access to 62 models and all of them chat completion models.

If so, I Don't know how they’ll provide the speech, vision or any other service they mentioned in the bottom page of “AAS”.

But the funny thing is if we create a resource in AI foundry service separately(let's call it “AIDRY”)it will fall under this Azure AI service(child), and will be available alongside AAS.

And if you try to open “AIDRY” in AI foundry from Azure AI service(child) it will have a slightly different button then AAS to open the foundry. And will offer all the 1800 models of all kinds.

Azure open AI

Creation : service/resource is created under the Azure -> Azure Ai Services -> Azure OpenAi

  Just creates a Open Service resource, then allows you to pick any model for any functionality 
  

Azure AI foundry

Creation : service/resource is created under the Azure -> Azure AI Foundry. Let's call this created resource “AAF”.

  This will take you directly to the foundry, where you can again create an Azure AI service(child).
  
     But the difference is you will have access to all the 1800+ models when you open the foundry.
     
        And this will automatically have the resources like speech and vision which are created along the way just like an Azure Ai service would do.
        

Lastely
questions:
1 why is it only 62 models and only chat completions.
2 what models will an Azure Ai service use when i'll try to call those vision or speech services using api??
3 Hubs(in ai foundry) i didn't get the concept all together what power do they hold which cannot be controlled by the AI foundry project itself

Answer 1

Hi Aditya patil

1. Restricted Access to Models: Azure AI services provide access to a variety of models, including OpenAI models and others from the Azure AI Foundry. However, the availability of models can depend on factors like region, subscription type, and specific service configurations. While there are over 1800 models in the Azure AI Model Catalog, not all may be accessible due to licensing, deployment limitations, or regional restrictions2. You might want to check the Azure AI Model Catalog for details on supported models and their availability.
2. Resource Duplication in AI Foundry: When you create a resource in Azure AI Foundry, it appears under Azure AI Services. This is by design, as Foundry resources are part of the broader Azure AI Services ecosystem. While it might seem redundant, this structure ensures that all AI resources are centrally managed.
3. Naming Conflict: The issue of a "parent" and "child" service having the same name could indeed be confusing. This might be related to Azure's naming conventions or a potential bug. You can review the naming rules and restrictions for Azure resources to ensure compliance.

Kindly refer below link: standard-chat-completions

Thank You.

Answer 2

hey @Saideep Anchuri i believe there was a misunderstanding.

1. I wanted to know when we open a AI foundry from the Azure AI service > Azure AI service > resource. I only get access to 62 only chat models.
2. In Azure AI resource creation, I directly receive a vision and speech model APIs. even when I haven't deployed any model, where are my requests being processed at and by which model????
3. when I open a Ai Foundry as an individual resource I get all the 1800+ models.
4. And when I open Ai Foundry through the Azure Open AI resource, I see I have only access to Open AI models.