Merge pull request #5567 from MicrosoftDocs/main

6/17/2025 AM Publish

Author: Taojunshen

articles/ai-services/openai/how-to/predicted-outputs.md

@@ -6,7 +6,7 @@ services: cognitive-services
 manager: nitinme
 ms.service: azure-ai-openai
 ms.topic: how-to
-ms.date: 04/14/2025
+ms.date: 06/17/2025
 author: mrbullwinkle
 ms.author: mbullwin
 recommendations: false
@@ -22,10 +22,12 @@ Predicted outputs can improve model response latency for chat completions calls
 - `gpt-4o` version: `2024-08-06`
 - `gpt-4o` version: `2024-11-20`
 - `gpt-4.1` version: `2025-04-14`
+- `gpt-4.1-nano` version: `2025-04-14`
+- `gpt-4.1-mini` version: `2025-04-14`
 
 ## API support
 
-- `2025-01-01-preview`
+First introduced in `2025-01-01-preview`. Supported in all subsequent releases. 
 
 ## Unsupported features
 

articles/ai-services/speech-service/includes/quickstarts/voice-live-api/realtime-python.md

@@ -159,8 +159,8 @@ For the recommended keyless authentication with Microsoft Entra ID, you need to:
                         "remove_filler_words": False,
                         "end_of_utterance_detection": {
                             "model": "semantic_detection_v1",
-                            "threshold": 0.1,
-                            "timeout": 4,
+                            "threshold": 0.01,
+                            "timeout": 2,
                         },
                     },
                     "input_audio_noise_reduction": {

articles/ai-services/speech-service/voice-live-how-to.md

@@ -16,7 +16,7 @@ ms.custom: references_regions
 
 [!INCLUDE [Feature preview](./includes/previews/preview-generic.md)]
 
-The Voice Live API provides a capable WebSocket interface compared to the [Azure OpenAI Realtime API](../openai/how-to/realtime-audio.md). 
+The Voice Live API provides a capable WebSocket interface compared to the [Azure OpenAI Realtime API](../openai/how-to/realtime-audio.md).
 
 Unless otherwise noted, the Voice Live API uses the same events as the [Azure OpenAI Realtime API](/azure/ai-services/openai/realtime-audio-reference?context=/azure/ai-services/speech-service/context/context). This document provides a reference for the event message properties that are specific to the Voice Live API.
 
@@ -26,7 +26,7 @@ For a table of supported models and regions, see the [Voice Live API overview](.
 
 ## Authentication
 
-An [Azure AI Foundry resource](../multi-service-resource.md) is required to access the Voice Live API. 
+An [Azure AI Foundry resource](../multi-service-resource.md) is required to access the Voice Live API.
 
 ### WebSocket endpoint
 
@@ -66,8 +66,8 @@ Here's an example `session.update` message that configures several aspects of th
         "remove_filler_words": false,
         "end_of_utterance_detection": {
             "model": "semantic_detection_v1",
-            "threshold": 0.1,
-            "timeout": 4,
+            "threshold": 0.01,
+            "timeout": 2,
         },
     },
     "input_audio_noise_reduction": {"type": "azure_deep_noise_suppression"},
@@ -84,10 +84,10 @@ The server responds with a [`session.updated`](../openai/realtime-audio-referenc
 
 ## Session Properties
 
-The following sections describe the properties of the `session` object that can be configured in the `session.update` message. 
+The following sections describe the properties of the `session` object that can be configured in the `session.update` message.
 
 > [!TIP]
-> For comprehensive descriptions of supported events and properties, see the [Azure OpenAI Realtime API events reference documentation](../openai/realtime-audio-reference.md?context=/azure/ai-services/speech-service/context/context). This document provides a reference for the event message properties that are enhancements via the Voice Live API. 
+> For comprehensive descriptions of supported events and properties, see the [Azure OpenAI Realtime API events reference documentation](../openai/realtime-audio-reference.md?context=/azure/ai-services/speech-service/context/context). This document provides a reference for the event message properties that are enhancements via the Voice Live API.
 
 ### Input audio properties
 
@@ -99,7 +99,7 @@ You can use input audio properties to configure the input audio stream.
 | `input_audio_echo_cancellation` | object   | Optional | Enhances the input audio quality by removing the echo from the model's own voice without requiring any client-side echo cancellation.<br/><br/>Set the `type` property of `input_audio_echo_cancellation` to enable echo cancellation.<br/><br/>The supported value for `type` is `server_echo_cancellation` which is used when the model's voice is played back to the end-user through a speaker, and the microphone picks up the model's own voice.  |
 | `input_audio_noise_reduction`   | object   | Optional | Enhances the input audio quality by suppressing or removing environmental background noise.<br/><br/>Set the `type` property of `input_audio_noise_reduction` to enable noise suppression.<br/><br/>The supported value for `type` is `azure_deep_noise_suppression` which optimizes for speakers closest to the microphone. |
 
-Here's an example of input audio properties is a session object: 
+Here's an example of input audio properties is a session object:
 
 ```json
 {
@@ -137,15 +137,15 @@ The Voice Live API offers conversational enhancements to provide robustness to t
 
 ### Turn Detection Parameters
 
-Turn detection is the process of detecting when the end-user started or stopped speaking. The Voice Live API builds on the Azure OpenAI Realtime API `turn_detection` property to configure turn detection. The `azure_semantic_vad` type is one differentiator between the Voice Live API and the Azure OpenAI Realtime API. 
+Turn detection is the process of detecting when the end-user started or stopped speaking. The Voice Live API builds on the Azure OpenAI Realtime API `turn_detection` property to configure turn detection. The `azure_semantic_vad` type is one differentiator between the Voice Live API and the Azure OpenAI Realtime API.
 
 | Property | Type | Required or optional | Description |
 |----------|----------|----------|------------|
 | `type` | string   | Optional | The type of turn detection system to use. Type `server_vad` detects start and end of speech based on audio volume.<br/><br/>Type `azure_semantic_vad` detects start and end of speech based on semantic meaning. Azure semantic voice activity detection (VAD) improves turn detection by removing filler words to reduce the false alarm rate. The current list of filler words are `['ah', 'umm', 'mm', 'uh', 'huh', 'oh', 'yeah', 'hmm']`. The service ignores these words when there's an ongoing response. Remove feature words feature assumes the client plays response audio as soon as it receives them. The `azure_semantic_vad` type isn't supported with the `gpt-4o-realtime-preview` and `gpt-4o-mini-realtime-preview` models.<br/><br/>The default value is `server_vad`. |
 | `threshold` | number | Optional | A higher threshold requires a higher confidence signal of the user trying to speak. |
 | `prefix_padding_ms` | integer | Optional  | The amount of audio, measured in milliseconds, to include before the start of speech detection signal. |
 | `silence_duration_ms` | integer  | Optional | The duration of user's silence, measured in milliseconds, to detect the end of speech. |
-| `end_of_utterance_detection` | object | Optional | Configuration for end of utterance detection. The Voice Live API offers advanced end-of-turn detection to indicate when the end-user stopped speaking while allowing for natural pauses. End of utterance detection can significantly reduce premature end-of-turn signals without adding user-perceivable latency. End of utterance detection is only available when using `azure_semantic_vad`.<br/><br/>Properties of `end_of_utterance_detection` include:<br/>-`model`: The model to use for end of utterance detection. The supported value is `semantic_detection_v1`.<br/>- `threshold`: Threshold to determine the end of utterance (0.0 to 1.0). The default value is 0.1.<br/>- `timeout`: Timeout in seconds. The default value is 4 seconds.|
+| `end_of_utterance_detection` | object | Optional | Configuration for end of utterance detection. The Voice Live API offers advanced end-of-turn detection to indicate when the end-user stopped speaking while allowing for natural pauses. End of utterance detection can significantly reduce premature end-of-turn signals without adding user-perceivable latency. End of utterance detection is only available when using `azure_semantic_vad`.<br/><br/>Properties of `end_of_utterance_detection` include:<br/>-`model`: The model to use for end of utterance detection. The supported value is `semantic_detection_v1`.<br/>- `threshold`: Threshold to determine the end of utterance (0.0 to 1.0). The default value is 0.01.<br/>- `timeout`: Timeout in seconds. The default value is 2 seconds.|
 
 Here's an example of end of utterance detection in a session object:
 
@@ -160,8 +160,8 @@ Here's an example of end of utterance detection in a session object:
             "remove_filler_words": false,
             "end_of_utterance_detection": {
                 "model": "semantic_detection_v1",
-                "threshold": 0.1,
-                "timeout": 4
+                "threshold": 0.01,
+                "timeout": 2
             }
         }
     }
@@ -170,7 +170,7 @@ Here's an example of end of utterance detection in a session object:
 
 ### Audio output through Azure text to speech
 
-You can use the `voice` parameter to specify a standard or custom voice. The voice is used for audio output. 
+You can use the `voice` parameter to specify a standard or custom voice. The voice is used for audio output.
 
 The `voice` object has the following properties:
 
@@ -357,7 +357,7 @@ To configure the viseme, you can set the `animation.outputs` in the `session.upd
 }
 ```
 
-The `output_audio_timestamp_types` parameter is optional. It configures which audio timestamps should be returned for generated audio. Currently, it only supports `word`. 
+The `output_audio_timestamp_types` parameter is optional. It configures which audio timestamps should be returned for generated audio. Currently, it only supports `word`.
 
 The service returns the viseme alignment in the response when the audio is generated.
 

articles/machine-learning/toc.yml

@@ -514,6 +514,9 @@ items:
   - name: Configure & submit training run
     displayName: run config, script run config, scriptrunconfig, compute target, dsvm, Data Science Virtual Machine, local, cluster, ACI, container instance, Databricks, data lake, lake, HDI, HDInsight
     href: ./v1/how-to-set-up-training-targets.md
+  - name: Log metrics, parameters, and files
+    displayName: troubleshoot, log, files, tracing, metrics
+    href: ./v1/how-to-log-view-metrics.md
   # end v1
   - name: Training with CLI and SDK
     href: how-to-train-model.md