Merge pull request #5133 from MicrosoftDocs/main

Merge main to live, 4 AM

Author: v-ccolin

.openpublishing.redirection.json

@@ -294,6 +294,21 @@
       "source_path_from_root": "/articles/ai-services/speech-service/text-to-speech-avatar/custom-avatar-endpoint.md",
       "redirect_url": "/azure/ai-services/speech-service/custom-avatar-create",
       "redirect_document_id": false
+    },
+    {
+      "source_path_from_root": "/articles/ai-services/speech-service/migration-overview-neural-voice.md",
+      "redirect_url": "/azure/ai-services/speech-service/custom-neural-voice",
+      "redirect_document_id": false
+    },
+    {
+      "source_path_from_root": "/articles/ai-services/speech-service/how-to-migrate-to-custom-neural-voice.md",
+      "redirect_url": "/azure/ai-services/speech-service/custom-neural-voice",
+      "redirect_document_id": false
+    },
+    {
+      "source_path_from_root": "/articles/ai-services/speech-service/how-to-migrate-to-prebuilt-neural-voice.md",
+      "redirect_url": "/azure/ai-services/speech-service/custom-neural-voice",
+      "redirect_document_id": false
     }
   ]
 }

articles/ai-foundry/how-to/develop/langchain.md

@@ -51,8 +51,11 @@ To use LLMs deployed in Azure AI Foundry portal, you need the endpoint and crede
 [!INCLUDE [tip-left-pane](../../includes/tip-left-pane.md)]
 
 1. Go to the [Azure AI Foundry](https://ai.azure.com/).
+
 1. Open the project where the model is deployed, if it isn't already open.
+
 1. Go to **Models + endpoints** and select the model you deployed as indicated in the prerequisites.
+
 1. Copy the endpoint URL and the key.
 
     :::image type="content" source="../../media/how-to/inference/serverless-endpoint-url-keys.png" alt-text="Screenshot of the option to copy endpoint URI and keys from an endpoint." lightbox="../../media/how-to/inference/serverless-endpoint-url-keys.png":::
@@ -63,11 +66,19 @@ To use LLMs deployed in Azure AI Foundry portal, you need the endpoint and crede
 In this scenario, we placed both the endpoint URL and key in the following environment variables:
 
 ```bash
-export AZURE_INFERENCE_ENDPOINT="<your-model-endpoint-goes-here>"
+export AZURE_INFERENCE_ENDPOINT="https://<resource>.services.ai.azure.com/models"
 export AZURE_INFERENCE_CREDENTIAL="<your-key-goes-here>"
 ```
 
-Once configured, create a client to connect to the endpoint. In this case, we're working with a chat completions model hence we import the class `AzureAIChatCompletionsModel`.
+Once configured, create a client to connect with the chat model by using the `init_chat_model`. For Azure OpenAI models, configure the client as indicated at [Using Azure OpenAI models](#using-azure-openai-models).
+
+```python
+from langchain.chat_models import init_chat_model
+
+llm = init_chat_model(model="mistral-large-2407", model_provider="azure_ai")
+```
+
+You can also use the class `AzureAIChatCompletionsModel` directly.
 
 ```python
 import os
@@ -80,8 +91,8 @@ model = AzureAIChatCompletionsModel(
 )
 ```
 
-> [!TIP]
-> For Azure OpenAI models, configure the client as indicated at [Using Azure OpenAI models](#using-azure-openai-models).
+> [!CAUTION]
+> **Breaking change:** Parameter `model_name` was renamed `model` in version `0.1.3`.
 
 You can use the following code to create the client if your endpoint supports Microsoft Entra ID:
 
@@ -93,7 +104,7 @@ from langchain_azure_ai.chat_models import AzureAIChatCompletionsModel
 model = AzureAIChatCompletionsModel(
     endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
     credential=DefaultAzureCredential(),
-    model_name="mistral-large-2407",
+    model="mistral-large-2407",
 )
 ```
 
@@ -111,7 +122,7 @@ from langchain_azure_ai.chat_models import AzureAIChatCompletionsModel
 model = AzureAIChatCompletionsModel(
     endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
     credential=DefaultAzureCredentialAsync(),
-    model_name="mistral-large-2407",
+    model="mistral-large-2407",
 )
 ```
 
@@ -188,13 +199,13 @@ from langchain_azure_ai.chat_models import AzureAIChatCompletionsModel
 producer = AzureAIChatCompletionsModel(
     endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
     credential=os.environ["AZURE_INFERENCE_CREDENTIAL"],
-    model_name="mistral-large-2407",
+    model="mistral-large-2407",
 )
 
 verifier = AzureAIChatCompletionsModel(
     endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
     credential=os.environ["AZURE_INFERENCE_CREDENTIAL"],
-    model_name="mistral-small",
+    model="mistral-small",
 )
 ```
 
@@ -271,7 +282,7 @@ from langchain_azure_ai.embeddings import AzureAIEmbeddingsModel
 embed_model = AzureAIEmbeddingsModel(
     endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
     credential=os.environ['AZURE_INFERENCE_CREDENTIAL'],
-    model_name="text-embedding-3-large",
+    model="text-embedding-3-large",
 )
 ```
 
@@ -305,31 +316,15 @@ for doc in results:
 
 ## Using Azure OpenAI models
 
-If you're using Azure OpenAI in Foundry Models or Foundry Models service with OpenAI models with `langchain-azure-ai` package, you might need to use `api_version` parameter to select a specific API version. The following example shows how to connect to an Azure OpenAI in Foundry Models deployment:
-
-```python
-from langchain_azure_ai.chat_models import AzureAIChatCompletionsModel
-
-llm = AzureAIChatCompletionsModel(
-    endpoint="https://<resource>.openai.azure.com/openai/deployments/<deployment-name>",
-    credential=os.environ["AZURE_INFERENCE_CREDENTIAL"],
-    api_version="2024-05-01-preview",
-)
-```
-
-> [!IMPORTANT]
-> Check which is the API version that your deployment is using. Using a wrong `api_version` or one not supported by the model results in a `ResourceNotFound` exception.
-
-If the deployment is hosted in Azure AI Services, you can use the Foundry Models service:
+If you're using Azure OpenAI models with `langchain-azure-ai` package, use the following URL:
 
 ```python
 from langchain_azure_ai.chat_models import AzureAIChatCompletionsModel
 
 llm = AzureAIChatCompletionsModel(
-    endpoint="https://<resource>.services.ai.azure.com/models",
+    endpoint="https://<resource>.openai.azure.com/openai/v1",
     credential=os.environ["AZURE_INFERENCE_CREDENTIAL"],
-    model_name="<model-name>",
-    api_version="2024-05-01-preview",
+    model="gpt-4o"
 )
 ```
 
@@ -370,7 +365,7 @@ from langchain_azure_ai.chat_models import AzureAIChatCompletionsModel
 model = AzureAIChatCompletionsModel(
     endpoint=os.environ["AZURE_INFERENCE_ENDPOINT"],
     credential=os.environ["AZURE_INFERENCE_CREDENTIAL"],
-    model_name="mistral-large-2407",
+    model="mistral-large-2407",
     client_kwargs={"logging_enable": True},
 )
 ```

articles/ai-foundry/model-inference/concepts/models.md

@@ -55,9 +55,9 @@ DeepSeek family of models includes DeepSeek-R1, which excels at reasoning tasks
 
 | Model  | Type | Tier | Capabilities |
 | ------ | ---- | ---- | ------------ |
-| [DeekSeek-V3-0324](https://ai.azure.com/explore/models/deepseek-v3-0324/version/1/registry/azureml-deepseek) | chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:**  (131,072 tokens) <br /> - **Languages:** `en` and `zh` <br />  - **Tool calling:** Yes <br /> - **Response formats:** Text, JSON |
-| [DeekSeek-R1](https://ai.azure.com/explore/models/deepseek-r1/version/1/registry/azureml-deepseek) | chat-completion <br /> [(with reasoning content)](../how-to/use-chat-reasoning.md) | Global standard | - **Input:** text (163,840 tokens) <br /> - **Output:**  (163,840 tokens) <br /> - **Languages:** `en` and `zh` <br />  - **Tool calling:** No <br /> - **Response formats:** Text. |
-| [DeekSeek-V3](https://ai.azure.com/explore/models/deepseek-v3/version/1/registry/azureml-deepseek) <br />(Legacy) | chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:**  (131,072 tokens) <br /> - **Languages:** `en` and `zh` <br />  - **Tool calling:** No <br /> - **Response formats:** Text, JSON |
+| [DeepSeek-V3-0324](https://ai.azure.com/explore/models/deepseek-v3-0324/version/1/registry/azureml-deepseek) | chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:**  (131,072 tokens) <br /> - **Languages:** `en` and `zh` <br />  - **Tool calling:** Yes <br /> - **Response formats:** Text, JSON |
+| [DeepSeek-R1](https://ai.azure.com/explore/models/deepseek-r1/version/1/registry/azureml-deepseek) | chat-completion <br /> [(with reasoning content)](../how-to/use-chat-reasoning.md) | Global standard | - **Input:** text (163,840 tokens) <br /> - **Output:**  (163,840 tokens) <br /> - **Languages:** `en` and `zh` <br />  - **Tool calling:** No <br /> - **Response formats:** Text. |
+| [DeepSeek-V3](https://ai.azure.com/explore/models/deepseek-v3/version/1/registry/azureml-deepseek) <br />(Legacy) | chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:**  (131,072 tokens) <br /> - **Languages:** `en` and `zh` <br />  - **Tool calling:** No <br /> - **Response formats:** Text, JSON |
 
 For a tutorial on DeepSeek-R1, see [Tutorial: Get started with DeepSeek-R1 reasoning model in Azure AI Foundry Models](../tutorials/get-started-deepseek-r1.md).
 

articles/ai-services/speech-service/batch-synthesis-properties.md

@@ -34,7 +34,7 @@ Batch synthesis properties are described in the following table.
 |`outputs.result`|The location of the batch synthesis result files with audio output and logs.<br/><br/>This property is read-only.|
 |`properties`|A defined set of optional batch synthesis configuration settings.|
 |`properties.sizeInBytes`|The audio output size in bytes.<br/><br/>This property is read-only.|
-|`properties.billingDetails`|The number of words that were processed and billed by `customNeuralCharacters` versus `neuralCharacters` (prebuilt) voices.<br/><br/>This property is read-only.|
+|`properties.billingDetails`|The number of words that were processed and billed by `customNeuralCharacters` (custom voice) versus `neuralCharacters` (standard voice).<br/><br/>This property is read-only.|
 |`properties.concatenateResult`|Determines whether to concatenate the result. This optional `bool` value ("true" or "false") is "false" by default.|
 |`properties.decompressOutputFiles`|Determines whether to unzip the synthesis result files in the destination container. This property can only be set when the `destinationContainerUrl` property is set. This optional `bool` value ("true" or "false") is "false" by default.|
 |`properties.destinationContainerUrl`|The batch synthesis results can be stored in a writable Azure container. If you don't specify a container URI with [shared access signatures (SAS)](/azure/storage/common/storage-sas-overview) token, the Speech service stores the results in a container managed by Microsoft. SAS with stored access policies isn't supported. When the synthesis job is deleted, the result data is also deleted.<br/><br/>This optional property isn't included in the response when you get the synthesis job.|
@@ -59,7 +59,7 @@ Batch synthesis properties are described in the following table.
 |`synthesisConfig.speakerProfileId`|The speaker profile ID of a personal voice.<br/><br/>For information about available personal voice base model names, see [integrate personal voice](personal-voice-how-to-use.md#integrate-personal-voice-in-your-application).<br/>For information about how to get the speaker profile ID, see [language and voice support](personal-voice-create-voice.md).<br/><br/>This property is required when `inputKind` is set to `"PlainText"`.|
 |`synthesisConfig.style`|For some voices, you can adjust the speaking style to express different emotions like cheerfulness, empathy, and calm. You can optimize the voice for different scenarios like customer service, newscast, and voice assistant.<br/><br/>For information about the available styles per voice, see [voice styles and roles](language-support.md?tabs=tts#voice-styles-and-roles).<br/><br/>This optional property is only applicable when `synthesisConfig.style` is set.|
 |`synthesisConfig.styleDegree`|The intensity of the speaking style. You can specify a stronger or softer style to make the speech more expressive or subdued. The range of accepted values are: 0.01 to 2 inclusive. The default value is 1, which means the predefined style intensity. The minimum unit is 0.01, which results in a slight tendency for the target style. A value of 2 results in a doubling of the default style intensity. If the style degree is missing or isn't supported for your voice, this attribute is ignored.<br/><br/>For information about the available styles per voice, see [voice styles and roles](language-support.md?tabs=tts#voice-styles-and-roles).<br/><br/>This optional property is only applicable when `inputKind` is set to `"PlainText"`.|
-|`synthesisConfig.voice`|The voice that speaks the audio output.<br/><br/>For information about the available prebuilt neural voices, see [language and voice support](language-support.md?tabs=tts). To use a custom voice, you must specify a valid custom voice and deployment ID mapping in the `customVoices` property. To use a personal voice, you need to specify the `synthesisConfig.speakerProfileId` property. <br/><br/>This property is required when `inputKind` is set to `"PlainText"`.|
+|`synthesisConfig.voice`|The voice that speaks the audio output.<br/><br/>For information about the available standard voices, see [language and voice support](language-support.md?tabs=tts). To use a custom voice, you must specify a valid custom voice and deployment ID mapping in the `customVoices` property. To use a personal voice, you need to specify the `synthesisConfig.speakerProfileId` property. <br/><br/>This property is required when `inputKind` is set to `"PlainText"`.|
 |`synthesisConfig.volume`|The volume of the audio output.<br/><br/>For information about the accepted values, see the [adjust prosody](speech-synthesis-markup-voice.md#adjust-prosody) table in the Speech Synthesis Markup Language (SSML) documentation. Invalid values are ignored.<br/><br/>This optional property is only applicable when `inputKind` is set to `"PlainText"`.|
 |`inputKind`|Indicates whether the `inputs` text property should be plain text or SSML. The possible case-insensitive values are "PlainText" and "SSML". When the `inputKind` is set to `"PlainText"`, you must also set the `synthesisConfig` voice property.<br/><br/>This property is required.|
 

articles/ai-services/speech-service/call-center-overview.md

@@ -42,7 +42,7 @@ The Speech service offers the following features that can be used for call cente
 - [Speaker identification](./speaker-recognition-overview.md): Helps you determine an unknown speaker’s identity within a group of enrolled speakers and is typically used for call center customer verification scenarios or fraud detection.
 - [Language Identification](./language-identification.md): Identify languages spoken in audio and can be used in real-time and post-call analysis for insights or to control the environment (such as output language of a virtual agent).
 
-The Speech service works well with prebuilt models. However, you might want to further customize and tune the experience for your product or environment. Typical examples for Speech customization include:
+You might want to further customize and fine-tune the experience for your product or environment. Typical examples for Speech fine-tuning include:
 
 | Speech customization | Description |
 | -------------- | ----------- |
@@ -57,7 +57,7 @@ The Language service offers the following features that can be used for call cen
 - [Conversation summarization](../language-service/summarization/overview.md?tabs=conversation-summarization): Summarize in abstract text what each conversation participant said about the issues and resolutions. For example, a call center can group product issues that have a high volume.
 - [Sentiment analysis and opinion mining](../language-service/sentiment-opinion-mining/overview.md): Analyze transcriptions and associate positive, neutral, or negative sentiment at the utterance and conversation-level.
 
-While the Language service works well with prebuilt models, you might want to further customize and tune models to extract more information from your data. Typical examples for Language customization include:
+You might want to further customize and fine-tune models to extract more information from your data. Typical examples for Language customization include:
 
 | Language customization | Description |
 | -------------- | ----------- |

articles/ai-services/speech-service/custom-neural-voice.md

@@ -19,7 +19,7 @@ Custom neural voice (CNV) is a text to speech feature that lets you create a one
 > 
 > Access to [Custom neural voice (CNV) Lite](custom-neural-voice-lite.md) is available for anyone to demo and evaluate CNV before investing in professional recordings to create a higher-quality voice. 
 
-Out of the box, [text to speech](text-to-speech.md) can be used with prebuilt neural voices for each [supported language](language-support.md?tabs=tts). The prebuilt neural voices work well in most text to speech scenarios if a unique voice isn't required.
+Out of the box, [text to speech](text-to-speech.md) can be used with standard voices for each [supported language](language-support.md?tabs=tts). The standard voices work well in most text to speech scenarios if a unique voice isn't required.
 
 Custom neural voice is based on the neural text to speech technology and the multilingual, multi-speaker, universal model. You can create synthetic voices that are rich in speaking styles, or adaptable cross languages. The realistic and natural sounding voice of custom neural voice can represent brands, personify machines, and allow users to interact with applications conversationally. See the [supported languages](language-support.md?tabs=tts) for custom neural voice.
 
@@ -46,7 +46,7 @@ Here's an overview of the steps to create a custom neural voice in Speech Studio
 1. [Test your voice](professional-voice-train-voice.md#test-your-voice-model). Prepare test scripts for your voice model that cover the different use cases for your apps. It’s a good idea to use scripts within and outside the training dataset, so you can test the quality more broadly for different content.
 1. [Deploy and use your voice model](professional-voice-deploy-endpoint.md) in your apps.
 
-You can tune, adjust, and use your custom voice, similarly as you would use a prebuilt neural voice. Convert text into speech in real-time, or generate audio content offline with text input. You use the [REST API](./rest-text-to-speech.md), the [Speech SDK](./get-started-text-to-speech.md), or the [Speech Studio](https://speech.microsoft.com/audiocontentcreation).
+You can tune, adjust, and use your custom voice, similarly as you would use a standard voice. Convert text into speech in real-time, or generate audio content offline with text input. You use the [REST API](./rest-text-to-speech.md), the [Speech SDK](./get-started-text-to-speech.md), or the [Speech Studio](https://speech.microsoft.com/audiocontentcreation).
 
 > [!TIP]
 > Check out the code samples in the [Speech SDK repository on GitHub](https://github.com/Azure-Samples/cognitive-services-speech-sdk/blob/master/samples/custom-voice/README.md) to see how to use custom neural voice in your application.