All Products
Search
Document Center

Platform For AI:Export labeling results

Last Updated:Apr 08, 2026

Export labeling results to an OSS bucket or create a dataset for model training.

Prerequisites

The labeling job is complete. For more information, see Process labeling jobs.

Limitations

Only an administrator or a labeling administrator can perform labeling operations. If you do not have the required permissions, ask an administrator to grant you the labeling administrator role. For more information, see Create and manage workspaces.

Export labeling results

  1. Go to iTAG.

    1. Log on to the PAI console.

    2. In the left navigation pane, click Workspaces. On the Workspaces page, click the workspace name.

    3. In the left navigation pane, choose Data Preparation > iTAGData Preparation.

  2. On the iTAG page, click the Jobs tab. Find the labeling job and click Export Labeling Result in the Actions column.

  3. In the Export Labeling Result dialog box, configure the parameters described in the following table and click OK.

    Export labeling results

    Parameter

    Description

    Export to Dataset

    Whether to create a dataset from labeling results for model training.

    To use the results for model training with PAI, select Yes.

    Output OSS Directory

    OSS path to store exported labeling results. If you also create a dataset, this path serves as the dataset storage path.

  4. After the export is complete, view the exported file.

    The exported file is in .manifest format. For file format details, see Labeling data format overview.

View export progress and results

  • In the upper-right corner of the iTAG page, click the Export progress icon to view export progress.

    For a successful export, click Redirect to open the OSS Bucket page and view the exported file. The file is in .manifest format. For format details, see Labeling data format overview.View export progress and results

  • If you created a dataset, view it on the Dataset Management page.Export and create a dataset

Next steps

Use the exported dataset for Designer, Data Science Workshop (DSW), or model training.