This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(275.2, 13%, 36%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ESV%3C/text%3E%3C/svg%3E)
Hi, We are trying to fine tune the "gpt-35-turbo" model programmatically using rest API with the resource in Sweden Central region. API url: https://xxxx.openai.azure.com/openai/fine-tunes?api-version=2023-07-01-preview Model details: { model = "gpt-35-turbo", training_file = "file-fileId" };
response = httpClient.PostAsync(url, httpContent).GetAwaiter().GetResult(); encountered the below error. "error": { "code": "invalidPayload", "message": "The specified base model does not support fine-tuning." } when I try to create a fine tune job with model as “gpt-35-turbo-0613”, getting the error as “The given base model is not supported on fine-tunes api route”. Please advise on how to fine tune "gpt-35-turbo" model Regards, Sunanda.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(22.400000000000002, 11%, 12%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ERM%3C/text%3E%3C/svg%3E)
Thanks for the question, Here is the document for models support fine-tuning via REST API:
gpt-35-turbo-0613
gpt-35-turbo-1106
Here is the models page to check which regions currently support fine-tuning.
Hello Sunanda Varma,
Regarding how to fine-tune GPT-35 Turbo, this Microsoft Learn article contains very detailed steps for reference.
Best regards,
Charlie
If you find my response helpful, please consider accepting this answer and voting 'yes' to support the community. Thank you!
Thanks, Ramr-msft,
I tried both gpt-35-turbo-0613 and gpt-35-turbo-1106 and received the below error message.
"The given base model is not supported on fine-tunes api route."
Sunanda.
Thanks for the reply, am able to create a custom fine tune model and also a deployment using this fine-tuned model from portal. So, are you saying Cognitive Services OpenAI Contributor assignment is needed only when fine tuning model using rest API?
Thanks for the details, you need Cognitive Services OpenAI Contributor assigned. Even someone with high-level Service Administrator permissions would still need this account explicitly set in order to access fine-tuning. For more information, please review the role-based access control guidance.
Thanks for the reply, am able to create a custom fine tune model and also a deployment using this fine-tuned model from portal. So, are you saying Cognitive Services OpenAI Contributor assignment is needed only when fine tuning model using rest API?
Thanks for the details, yes.
Seems the role is granted by default when the OpenAI instance is created. Please check below. Also, when calling REST API doesn't it use API key? why is the need for user to have the role assigned?
Thanks for the reply, am able to create a custom fine tune model and also a deployment using this fine-tuned model from portal. So, are you saying Cognitive Services OpenAI Contributor assignment is needed only when fine tuning model using rest API?
Yes, Thanks for the details.
Seems the role is granted by default when the OpenAI instance is created. Please check below. Also, when calling REST API doesn't it use API key? why is the need for user to have the role assigned?