Getting error from AzureOpenAI endpoint: logprobs, best_of and echo parameters are not available on gpt-35-turbo model

[hario90]

Describe the bug
Getting 400 HTTP status code response from AzureOpenAI when serviceId and deploymentOrModelId are set to "gpt-35-turbo".

fail: Microsoft.SemanticKernel.IKernel[0]
      Something went wrong while rendering the semantic function or while executing the text completion. Function: FunSkill.Joke. Error: Invalid request: The request is not valid, HTTP status: 400. Details: logprobs, best_of and echo parameters are not available on gpt-35-turbo model. Please remove the parameter and try again. For more details, see https://go.microsoft.com/fwlink/?linkid=2227346.
      Status: 400 (BadRequest)
      ErrorCode: BadRequest

      Content:
      {"error":{"code":"BadRequest","message":"logprobs, best_of and echo parameters are not available on gpt-35-turbo model. Please remove the parameter and try again. For more details, see https://go.microsoft.com/fwlink/?linkid=2227346."}}

To Reproduce
Steps to reproduce the behavior:

1. Clone https://github.com/microsoft/semantic-kernel-starters/tree/main/sk-csharp-console-chat
2. In terminal, change directory to sk-csharp-hello-world
3. follow README.md instructions in this directory to set dotnet secrets for serviceId and deploymentOrModelId to "gpt-35-turbo"
4. In terminal, run

dotnet build
dotnet run

Expected behavior
Console app runs without errors.

Screenshots
If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Windows
• IDE: N/A
• NuGet Package Version: 0.15.230531.5-preview and 0.16.230615.1-preview (reproducible in both)

Additional context
Add any other context about the problem here.

[hario90]

Closing since I was using a ChatCompletion model when the EndpointType in KernelSettings was set to TextCompletion.

[ratdoux]

Arghh thank you!!!

[CalvinQuark]

I encountered this issue while intentionally targeting the gpt-35-turbo model. This model is not one indicated in the Hello World example's README file which instead specifies that the text-davinci-003 model is to be targeted.

When I created a new Azure OpenAI resource, I requested the full gamut of capabilities in the approval request step. Nevertheless, the text-davinci-003 model is not deployable in my Azure AI Studio instance. The Studio portal indicates "No" in the Deployable column and provides a tooltip stating "This model must be fine-tuned before it can be deployed". But the Studio doesn't appear to either include directly or link to instructions regarding how the model can be fine-tuned and therefore deployed. Hence, the onboarding experience is a bit confusing to the uninitiated.

[hario90]

@CalvinQuark That's a good point, I will add a troubleshooting guide for this error and some information on how to update the endpoint type to the README.

[hario90]

TSGs can be found here: https://github.com/microsoft/semantic-kernel-starters/tree/main/sk-csharp-azure-functions#troubleshooting and here: https://github.com/microsoft/semantic-kernel-starters/tree/main/sk-csharp-hello-world#troubleshooting