Help Center/ PanguLargeModels/ User Guide/ Accessing Models in the Model Square
Updated on 2025-07-28 GMT+08:00
View PDF
Share

Accessing Models in the Model Square

ModelArts Studio provides a one-stop model resource library on the Model Square page. You can select a proper model to execute training and inference jobs based on service requirements.

In the left pane of the Model Square page, you can select different filter criteria to filter the required models, or search for models by keyword. The right pane displays the models in cards, including the model name, model description, and application scenario tags. You can click View Details to go to the model details page and view the model version, description, and details. You can also click Train or Deploy to go to the model training or deployment page and create a training or deployment job for the model.

For details about the large models supported by the platform, see "Model Capabilities and Specifications" > "Third-Party Large Models" in Service Overview.

Accessing the Model Square

  1. Log in to ModelArts Studio Large Model Deveopment Platform. In the My Spaces area, click the required workspace.
    Figure 1 My Spaces
  2. In the navigation pane, choose Model Square.
  3. In the left pane of the Model Square page, filter models by model type, model series, application scenario, supported function, and subscription type.
    Table 1 Model filtering

    Filter Criteria

    Description

    Model Type

    Model types include NLP, CV, AI4Science, etc.

    Model Series

    You can select Pangu Basic Large Model.

    Application Scenarios

    The options include LongTextInference, LogicalInference, SemanticSegmentation, ObjectDetection, DeepThinking, TextGeneration, DataAnalysis, IntentClassification, etc.

    Supported Functions

    The options include Support Training and Support Deployment.

    Subscription Type

    The options include Free and Paid.

Viewing Model Version Information

On the Model Square page, move the pointer over the target model card and click View Details. On the model details page that is displayed, view the model details.

Figure 2 Model Square

The model details page displays the model introduction and version list. The Model Introduction tab page displays the model description, disclaimer, and platform service agreement. The Version List tab page displays the features of multiple versions of the current model.

Figure 3 Model Introduction
Figure 4 Version List

Training and Deploying a Model

For details about how to train and deploy a model, see the sections of the corresponding model.

Trying Out Model Functions

Deploy an existing model. The platform provides preset deployment services. You can click Experience Center to try out the model functions. For models that do not provide preset services, you need to deploy them on the Model Development > Model Deployment page.

  1. In the navigation pane, choose Model Development > Model Deployment. Click Create Deployment Task in the upper right corner.
  2. On the displayed page, select the model source and model type, set deployment parameters, and start model deployment.

    After the deployment is complete, the Experience Now button is displayed on the card of the corresponding model on the Model Square page. After you click the button, the Experience Center page is displayed.