#Model deployment

11 messages · Page 1 of 1 (latest)

deft lion
#

Someone please help me with this error, I am stuck on this for days. I am just deploying a simple gpt-4o model. I recreated resources, hubs, projects all many times.

#

I once tried to create a connection with the name it was looking for but got

versed pilot
#

have you tried changing the name for "Connection Name"? The error is pretty clear

deft lion
#

Yes. When I am creating the project I try to create a connection named Default_AzureOpenAI but it doesn't let me as it does not accept underscores. When I try to create a connection later it gives me the error of it being a reserved connection name.

versed pilot
#

the connection name can be anything you want. Default_AzureOpenAI is a reserved name. Just enter another name.

deft lion
#

Yes tried that but then the error remains the same.

versed pilot
#

But I don't understand. Are you having trouble creating a deployment or a connection?

deft lion
#

The model deployment is giving the connection error.

#

If you see my first screen shot, I deployed gpt-4o but it gives 'Failed to get a connection error'. So, when I try to use the model in like prompt flow. It doesn't work. And then creating the connection has its own issues.

versed pilot
#

and your deployment is not usable without the Default_AzureOpenAI connection?

midnight brook
#

@deft lion please delete all you deployments this error is usually when you don't have quota capacity. Please recreate a new project and ensure it's in Sweden Central or France Central and try to deploy and new GPT4 model.