Merge pull request #233510 from mrbullwinkle/mrb_04_05_2023_fine_tuning_update

[Cognitive Services] [Azure OpenAI] Fine-tuning update

Author: AnnaMHuff

articles/cognitive-services/openai/concepts/models.md

@@ -250,4 +250,5 @@ These models can only be used with Embedding API requests.
 
 ## Next steps
 
-[Learn more about Azure OpenAI](../overview.md)
+- [Learn more about Azure OpenAI](../overview.md)
+- [Learn more about fine-tuning Azure OpenAI models](../how-to/fine-tuning.md)

articles/cognitive-services/openai/faq.yml

@@ -100,7 +100,10 @@ sections:
           What are the SLAs for API responses in Azure OpenAI?
         answer:
           We don't have a defined API response time Service Level Agreement (SLA) at this time. The overall SLA for Azure OpenAI Service is the same as for other Azure Cognitive Services. For more information, see the Cognitive Services section of the [Service Level Agreements (SLA) for Online Services page](https://azure.microsoft.com/support/legal/sla/cognitive-services/v1_1/).  
-  
+      - question: |
+          Why was my fine-tuned model deployment deleted? 
+        answer:
+          If a customized (fine-tuned) model is deployed for more than fifteen (15) days during which no completions or chat completions calls are made to it, the deployment will automatically be deleted (and no further hosting charges will be incurred for that deployment). The underlying customized model will remain available and can be redeployed at any time. To learn more check out the [how-to-article](/azure/cognitive-services/openai/how-to/fine-tuning?pivots=programming-language-studio#deploy-a-customized-model).  
 additionalContent: |
 
   ## Next steps

articles/cognitive-services/openai/how-to/fine-tuning.md

@@ -7,7 +7,7 @@ manager: nitinme
 ms.service: cognitive-services
 ms.subservice: openai
 ms.topic: how-to
-ms.date: 12/14/2022
+ms.date: 04/05/2023
 author: ChrisHMSFT
 ms.author: chrhoder
 zone_pivot_groups: openai-fine-tuning

articles/cognitive-services/openai/how-to/manage-costs.md

@@ -7,7 +7,7 @@ ms.custom: subject-cost-optimization
 ms.service: cognitive-services
 ms.subservice: openai
 ms.topic: how-to
-ms.date: 02/13/2023
+ms.date: 04/05/2023
 ---
 
 
@@ -49,6 +49,8 @@ Azure OpenAI fine-tuned models are charged based on three factors:
 
 The hosting hours cost is important to be aware of since once a fine-tuned model is deployed it continues to incur an hourly cost regardless of whether you're actively using it. Fine-tuned model costs should be monitored closely.
 
+[!INCLUDE [Fine-tuning deletion](../includes/fine-tune.md)]
+
 ### Other costs that might accrue with Azure OpenAI Service
 
 Keep in mind that enabling capabilities like sending data to Azure Monitor Logs, alerting, etc. incurs additional costs for those services. These costs are visible under those other services and at the subscription level, but aren't visible when scoped just to your Azure OpenAI resource.

articles/cognitive-services/openai/includes/fine-tune.md

@@ -0,0 +1,16 @@
+---
+title: Fine-tuning inactivity guidance
+titleSuffix: Azure OpenAI
+description: Fine-tuning inactivity guidance
+author: mrbullwinkle 
+ms.author: mbullwin 
+ms.service: cognitive-services
+ms.topic: include
+ms.date: 04/05/2023
+manager: nitinme
+keywords: ChatGPT
+
+---
+
+> [!IMPORTANT]
+> After a customized model is deployed, if at any time the deployment remains inactive for greater than fifteen (15) days, the deployment will automatically be deleted. The deployment of a customized model is “inactive” if the model was deployed more than fifteen (15) days ago and no completions or chat completions calls were made to it during a continuous 15-day period. The deletion of an inactive deployment does **NOT** delete or affect the underlying customized model, and the customized model can be redeployed at any time. As described in [Azure OpenAI Service pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/), each customized (fine-tuned) model that is deployed incurs an hourly hosting cost regardless of whether completions or chat completions calls are being made to the model. To learn more about planning and managing costs with Azure OpenAI, refer to our [cost management guide](/azure/cognitive-services/openai/how-to/manage-costs#base-series-and-codex-series-fine-tuned-models). 

articles/cognitive-services/openai/includes/fine-tuning-python.md

@@ -7,7 +7,7 @@ manager: nitinme
 ms.service: cognitive-services
 ms.subservice: openai
 ms.topic: how-to
-ms.date: 06/30/2022
+ms.date: 04/05/2023
 author: ChrisHMSFT
 ms.author: chrhoder
 keywords: 
@@ -237,6 +237,8 @@ print(f'Found {len(result)} fine-tune jobs.')
 
 When the fine-tune job has succeeded, the value of `fine_tuned_model` in the response body of the FineTune.retrieve() method is set to the name of your customized model. Your model is now also available for discovery from the [list Models API](/rest/api/cognitiveservices/azureopenaistable/models/list). However, you can't issue completion calls to your customized model until your customized model is deployed. You must deploy your customized model to make it available for use with completion calls.
 
+[!INCLUDE [Fine-tuning deletion](fine-tune.md)]
+
 > [!NOTE]
 > As with all applications, we require a review process prior to going live.
 
@@ -344,6 +346,8 @@ When you're done with your customized model, you can delete the deployment and m
 
 ### Delete your model deployment
 
+[!INCLUDE [Fine-tuning deletion](fine-tune.md)]
+
 You can use various methods to delete the deployment for your customized model:
 
 - [Azure OpenAI Studio](../how-to/fine-tuning.md?pivots=programming-language-studio#delete-your-model-deployment)</a>

articles/cognitive-services/openai/includes/fine-tuning-rest.md

@@ -245,6 +245,8 @@ print(f'Found {len((r.json())["data"])} fine-tune jobs.')
 
 When the fine-tune job has succeeded, the value of `fine_tuned_model` in the response body of the `FineTune.retrieve()` method is set to the name of your customized model. Your model is now also available for discovery from the [list Models API](/rest/api/cognitiveservices/azureopenaistable/models/list). However, you can't issue completion calls to your customized model until your customized model is deployed. You must deploy your customized model to make it available for use with completion calls.
 
+[!INCLUDE [Fine-tuning deletion](fine-tune.md)]
+
 > [!NOTE]
 > As with all applications, we require a review process prior to going live.
 
@@ -380,6 +382,8 @@ When you're done with your customized model, you can delete the deployment and m
 
 ### Delete your model deployment
 
+[!INCLUDE [Fine-tuning deletion](fine-tune.md)]
+
 You can use various methods to delete the deployment for your customized model:
 
 - [Azure OpenAI Studio](../how-to/fine-tuning.md?pivots=programming-language-studio#delete-your-model-deployment)

articles/cognitive-services/openai/includes/fine-tuning-studio.md

@@ -241,6 +241,8 @@ From the model page, you can also select **Download training file** to download
 
 When the fine-tune job has succeeded, you can deploy the customized model from the **Models** pane. You must deploy your customized model to make it available for use with completion calls.
 
+[!INCLUDE [Fine-tuning deletion](fine-tune.md)]
+
 > [!NOTE]
 > Only one deployment is permitted for a customized model. An error message is displayed if you select an already-deployed customized model.
 
@@ -287,6 +289,8 @@ When you're done with your customized model, you can delete the deployment and m
 
 ### Delete your model deployment
 
+[!INCLUDE [Fine-tuning deletion](fine-tune.md)]
+
 You can delete the deployment for your customized model from the **Deployments** page for Azure OpenAI Studio. Select the deployment to delete, and then select **Delete** to delete the deployment. 
 
 ### Delete your customized model

articles/cognitive-services/openai/whats-new.md

@@ -15,6 +15,11 @@ keywords:
 
 # What's new in Azure OpenAI Service
 
+## April 2023
+
+- **Inactive deployments of customized models will now be deleted after 15 days; models will remain available for redeployment.** If a customized (fine-tuned) model is deployed for more than fifteen (15) days during which no completions or chat completions calls are made to it, the deployment will automatically be deleted (and no further hosting charges will be incurred for that deployment). The underlying customized model will remain available and can be redeployed at any time. To learn more check out the [how-to-article](/azure/cognitive-services/openai/how-to/fine-tuning?pivots=programming-language-studio#deploy-a-customized-model).
+
+
 ## March 2023
 
 - **GPT-4 series models are now available in preview on Azure OpenAI**. To request access, existing Azure OpenAI customers can [apply by filling out this form](https://aka.ms/oai/get-gpt4). These models are currently available in the East US and South Central US regions.