You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
1. Sign in to [Azure AI Foundry](https://ai.azure.com/?cid=learnDocs).
18
-
1. Select any existing project if you aren't already in one.
19
-
20
-
> [!IMPORTANT]
21
-
> You must have an existing [!INCLUDE [fdp-project-name](fdp-project-name.md)] or [!INCLUDE [hub-project-name](hub-project-name.md)] before you can follow these steps. If you don't have an existing project, follow the steps in the [Quickstart: Get started with Azure AI Foundry](../quickstarts/get-started-code.md) to create your first [!INCLUDE [fdp-project-name](fdp-project-name.md)].
15
+
1. Sign in to [Azure AI Foundry](https://ai.azure.com/?cid=learnDocs).
22
16
23
-
1. At the bottom of the left pane, select **Management center**.
24
-
1. At the top of the page, select **All resources**.
25
-
1. Select **Create**.
26
-
27
-
:::image type="content" source="../media/how-to/projects/management-center-create.png" alt-text="Screenshot shows the management center where you can create a hub based project.":::
1. If you have a hub, you'll see the one you most recently used selected.
32
-
* If you have access to more than one hub, you can select a different hub from the dropdown.
33
-
* If you want to create a new one, select **Create new hub** and supply a name.
34
-
35
-
:::image type="content" source="../media/how-to/projects/projects-create-details.png" alt-text="Screenshot of the project details page within the create project dialog." lightbox="../media/how-to/projects/projects-create-details.png":::
36
-
37
-
1. If you don't have a hub, a default one is created for you.
38
-
1. Select **Create**.
19
+
1. Select **AI hub resource**, then select **Next**.
20
+
1. Enter a name for the project.
21
+
1. If you have a hub, you'll see the one you most recently used selected.
22
+
* If you have access to more than one hub, you can select a different hub from the dropdown.
23
+
* If you want to create a new one, select **Create new hub** and supply a name.
24
+
25
+
:::image type="content" source="../media/how-to/projects/projects-create-details.png" alt-text="Screenshot of the project details page within the create project dialog." lightbox="../media/how-to/projects/projects-create-details.png":::
26
+
27
+
1. If you don't have a hub, a default one is created for you.
***If you don't have any existing projects**: Follow the steps in [Quickstart: Get started with Azure AI Foundry](../quickstarts/get-started-code.md) to create your first project.
16
+
17
+
***If you're in a project**: Select the project breadcrumb, then select **Create new resource**.
18
+
19
+
:::image type="content" source="../media/how-to/projects/create-new-resource.png" alt-text="Screenshot shows creating a new project from a breadcrumb.":::
20
+
21
+
***If you're not in a project**: Select **Create new** in the top right to create a new [!INCLUDE [fdp-project-name-plural](fdp-project-name.md)]
22
+
23
+
:::image type="content" source="../media/how-to/projects/create-new.png" alt-text="Screenshot shows how to create a new project in Azure AI Foundry.":::
***If you don't have any existing projects**: Follow the steps in [Quickstart: Get started with Azure AI Foundry](../quickstarts/get-started-code.md) to create your first project.
71
-
72
-
***If you're in a project**: Select the project breadcrumb, then select **Create new resource**.
73
-
74
-
:::image type="content" source="../media/how-to/projects/create-new-resource.png" alt-text="Screenshot shows creating a new project from a breadcrumb.":::
75
-
76
-
***If you're not in a project**: Select **Create new** in the top right to create a new [!INCLUDE [fdp-project-name-plural](fdp-project-name.md)]
77
-
78
-
:::image type="content" source="../media/how-to/projects/create-new.png" alt-text="Screenshot shows how to create a new project in Azure AI Foundry.":::
70
+
1. Select **Azure AI Foundry resource**, then select **Next**.
71
+
1. Provide a name for your project and select **Create**. Or see next section for advanced options.
Copy file name to clipboardExpand all lines: articles/ai-foundry/model-inference/concepts/models.md
+10-9Lines changed: 10 additions & 9 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -97,6 +97,16 @@ Meta Llama models and tools are a collection of pretrained and fine-tuned genera
97
97
98
98
See [this model collection in Azure AI Foundry portal](https://ai.azure.com/explore/models?&selectedCollection=meta).
99
99
100
+
### xAI
101
+
102
+
xAI's Grok 3 and Grok 3 Mini models are designed to excel in various enterprise domains. Grok 3, a non-reasoning model pre-trained by the Colossus datacenter, is tailored for business use cases such as data extraction, coding, and text summarization, with exceptional instruction-following capabilities. It supports a 131,072 token context window, allowing it to handle extensive inputs while maintaining coherence and depth, and is particularly adept at drawing connections across domains and languages. On the other hand, Grok 3 Mini is a lightweight reasoning model trained to tackle agentic, coding, mathematical, and deep science problems with test-time compute. It also supports a 131,072 token context window for understanding codebases and enterprise documents, and excels at using tools to solve complex logical problems in novel environments, offering raw reasoning traces for user inspection with adjustable thinking budgets.
103
+
104
+
| Model | Type | Tier | Capabilities |
105
+
| ------ | ---- | --- | ------------ |
106
+
|[grok-3](https://ai.azure.com/explore/models/grok-3/version/1/registry/azureml-xai)| chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:** text (131,072 tokens) <br /> - **Languages:**`en` <br /> - **Tool calling:** yes <br /> - **Response formats:** text |
107
+
|[grok-3-mini](https://ai.azure.com/explore/models/grok-3-mini/version/1/registry/azureml-xai)| chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:** text (131,072 tokens) <br /> - **Languages:**`en` <br /> - **Tool calling:** yes <br /> - **Response formats:** text |
108
+
109
+
100
110
## Models from Partners and Community
101
111
102
112
Models from Partners and Community available for deployment with pay-as-you-go billing (for example, Cohere models) are offered by the model provider but hosted in Microsoft-managed Azure infrastructure and accessed via API in the Azure AI Foundry. Model providers define the license terms and set the price for use of their models, while Azure AI Foundry manages the hosting infrastructure.
@@ -210,15 +220,6 @@ See [this model collection in Azure AI Foundry portal](https://ai.azure.com/expl
210
220
| ------ | ---- | --- | ------------ |
211
221
|[tsuzumi-7b](https://ai.azure.com/explore/models/Tsuzumi-7b/version/1/registry/azureml-nttdata)| chat-completion | Global standard | - **Input:** text (8,192 tokens) <br /> - **Output:** text (8,192 tokens) <br /> - **Languages:**`en` and `jp` <br /> - **Tool calling:** No <br /> - **Response formats:** Text |
212
222
213
-
### xAI
214
-
215
-
xAI's Grok 3 and Grok 3 Mini models are designed to excel in various enterprise domains. Grok 3, a non-reasoning model pre-trained by the Colossus datacenter, is tailored for business use cases such as data extraction, coding, and text summarization, with exceptional instruction-following capabilities. It supports a 131,072 token context window, allowing it to handle extensive inputs while maintaining coherence and depth, and is particularly adept at drawing connections across domains and languages. On the other hand, Grok 3 Mini is a lightweight reasoning model trained to tackle agentic, coding, mathematical, and deep science problems with test-time compute. It also supports a 131,072 token context window for understanding codebases and enterprise documents, and excels at using tools to solve complex logical problems in novel environments, offering raw reasoning traces for user inspection with adjustable thinking budgets.
216
-
217
-
| Model | Type | Tier | Capabilities |
218
-
| ------ | ---- | --- | ------------ |
219
-
|[grok-3](https://ai.azure.com/explore/models/grok-3/version/1/registry/azureml-xai)| chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:** text (131,072 tokens) <br /> - **Languages:**`en` <br /> - **Tool calling:** yes <br /> - **Response formats:** text |
220
-
|[grok-3-mini](https://ai.azure.com/explore/models/grok-3-mini/version/1/registry/azureml-xai)| chat-completion | Global standard | - **Input:** text (131,072 tokens) <br /> - **Output:** text (131,072 tokens) <br /> - **Languages:**`en` <br /> - **Tool calling:** yes <br /> - **Response formats:** text |
221
-
222
223
## Open and protected models
223
224
224
225
The [Azure AI model catalog](../../../ai-studio/how-to/model-catalog-overview.md) offers a larger selection of models, from a bigger range of providers. As opposite to Azure AI Foundry Models where models are provided as APIs, these models might require you to host them on your infrastructure, including the creation of an AI hub and project, and providing the underlying compute quota to host the models.
> Precision, recall and F1 score are calculated for each class separately (*class-level* evaluation) and for the model collectively (*model-level* evaluation).
34
+
> Precision, recall, and F1 score are calculated for each class separately (*class-level* evaluation) and for the model collectively (*model-level* evaluation).
35
35
## Model-level and Class-level evaluation metrics
36
36
37
37
The definitions of precision, recall, and evaluation are the same for both class-level and model-level evaluations. However, the count of *True Positive*, *False Positive*, and *False Negative* differ as shown in the following example.
@@ -89,7 +89,7 @@ The below sections use the following example dataset:
> For single-label classification models, the count of false negatives and false positives are always equal. Custom single-label classification models always predict one class for each document. If the prediction is not correct, FP count of the predicted class increases by one and FN of the actual class increases by one, overall count of FP and FN for the model will always be equal. This is not the case for multi-label classification, because failing to predict one of the classes of a document is counted as a false negative.
92
+
> For single-label classification models, the number of false negatives and false positives are always equal. Custom single-label classification models always predict one class for each document. If the prediction is not correct, FP count of the predicted class increases by one and FN of the actual class increases by one, overall count of FP and FN for the model will always be equal. This is not the case for multi-label classification, because failing to predict one of the classes of a document is counted as a false negative.
93
93
## Interpreting class-level evaluation metrics
94
94
95
95
So what does it actually mean to have a high precision or a high recall for a certain class?
0 commit comments