Resource not found using client.get_model_info()

strawberrybfs 0

I was following this tutorial https://learn.microsoft.com/en-us/azure/ai-studio/how-to/deploy-models-llama?tabs=python-llama-3-2%2Cmeta-llama-3-1&pivots=programming-language-python

I used the following code and got error azure.core.exceptions.ResourceNotFoundError: (404) Resource not found

Does anyone know what happened here? Thank you!

client = ChatCompletionsClient(
    endpoint=api_llama.AZURE_ENDPOINT,
    credential=AzureKeyCredential(api_llama.API_KEY),
)

model_info = client.get_model_info()

1 answer

Max Lacy 180 Reputation points

2024-11-24T21:45:40.3366667+00:00
Hello strawberrybfs,

Looks like you're working through the tutorial for deploying models. Typically, when I have a "Resource not found" error I like to verify my API endpoint and key. You can use a debugger or print to your terminal with a print statement above the chat completions client.

print(endpoint, key) ## if you want to debug using print statements client = ChatCompletionsClient(endpoint=api_llama.AZURE_ENDPOINT, credential=AzureKeyCredential(api_llama.API_KEY), ) model_info = client.get_model_info()

Check that your env variables match the endpoints setup from the earlier steps in the tutorial. As a reminder to ensure the security of your applications and data, never share your API keys with anyone.

If they env variables are as you expect. You can use AI foundry for example code to compare your code.

An alternate route would be to use something like Postman to do a API call using the same tutorial but instead of using python select REST as the programming language.
Please sign in to rate this answer.
strawberrybfs 0 Reputation points

2024-11-24T22:48:40.7033333+00:00

Hi Max, thank you for the help! Yes I'm following the tutorial. My endpoint and key seem to not align with what's mentioned in the tutorial. Do you know where I can find my unique model deployment host name and the 32-character key? I got them from AI Foundry -> (sive nav) my assets -> Models + endpoints -> Endpoint -> Target URI + key

Is it correct?

Where to find your-host-name?

Where to find a key that's 32-character string? Mine is way longer than that.

strawberrybfs 0 Reputation points

2024-11-24T22:50:54.1166667+00:00

Hi, my previous comment seems disappearing so I'm posting it again. It seems like my endpoint and key is not aligned with what's mentioned in the tutorial. Do you know where I can find my unique model deployment host name and the 32-character key? I went to AI foundry -> my asset -> models + endpoints only finding a very long key and an endpoint. Thank you!

Max Lacy 180 Reputation points

2024-11-25T03:33:01.1266667+00:00

Hello,

Before we continue troubleshooting I want to confirm some assumptions I made with my answer.

Your using the Deployment to serverless APIs deployment, rather than the Deployment to a self-hosted managed compute.

Your intention is to use an API key for authentication.

You'd like to use local environment variables to store your API key.

If those assumptions are correct then following then following these steps should allow you to access your model:

Install the azure-ai-inference in your active environment. I'd recommend doing some sort of environment control. Learn about Python Environments

pip install azure-ai-inference

set your env variables based on what you found in AI Foundry -> (sive nav) my assets -> Models + endpoints -> Endpoint -> Target URI + key. This variable will only be available in the current CMD session. If you want it permanently available use setx rather than set. When using setx to set environment variables permanently on Windows, be cautious as it can affect all future sessions and might require a system restart to take effect. Additionally, improperly setting environment variables can introduce vulnerabilities.

set set

run your code with optional debugging to test your connection.

import os from azure.ai.inference import ChatCompletionsClient from azure.core.credentials import AzureKeyCredential client = ChatCompletionsClient( endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"], credential=AzureKeyCredential(os.environ["AZURE_INFERENCE_CREDENTIAL"]), ) model_info = client.get_model_info()

If this fails please share your model deployment and any configuration information.
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Share via

Resource not found using client.get_model_info()

1 answer

Your answer