Create annotation jobs using preset templates for image, text, video, and audio data, or build custom templates for specific scenarios.
Prerequisites
-
Activate PAI with a workspace. See Activate PAI and create a default workspace or Create and manage workspaces.
-
Activate OSS. Upload data files to OSS and create a dataset. See Create a dataset for data annotation.
Limitation
Only administrators or annotation administrators can perform labeling operations. Contact an administrator to grant annotation administrator permission if needed. See Manage workspace members.
Procedure
-
Go to the iTAG page.
-
Log on to the PAI console.
-
In the left-side navigation pane, click Workspaces. On the Workspaces page, click the name of the workspace that you want to manage.
-
In the left-side navigation pane, choose .
-
-
On the iTAG page, go to the Jobs tab and click Create Task.
-
On the Create Labeling Job page, configure the following parameters.
Parameter
Description
Task Name
1 to 100 characters. Must start with a letter, digit, or Chinese character. Can include underscores (_) or hyphens (-).
Input Dataset
Select a dataset from PAI dataset management.
Template Type
Select a template type:
-
General Template: Preset platform templates.
-
Custom Template: Build custom templates for specific scenarios. Combine Content Components and Topic Components as prompted.
For input and output data formats, see Custom templates.
Template
If Template Type is General Template, select a template:
-
Image
For application scenarios and data formats, see Image.
-
OCR: Recognize text within a selected image area.
-
Object Detection: Detect and locate objects in images.
-
Image Classification: Classify images with preset tags.
-
PDF: Perform OCR and classification on PDF files.
-
Moderation and Matting: Moderate and perform matting on images.
-
Table Recognition: Pre-recognize core table elements. Edit results as needed.
-
-
Text
For application scenarios and data formats, see Text.
-
Entity Recognition: Recognize named entities in text.
-
Text Classification: Classify text with preset tags. Supports single-tag and multi-tag classification.
-
Entity relationship: Identify relationships between text entities for knowledge graph scenarios.
-
-
Video
For application scenarios and data formats, see Video.
-
Video Classification: Classify videos with preset tags. Supports single-tag and multi-tag classification.
-
-
Audio
For application scenarios and data formats, see Audio.
-
Audio Classification: Classify audio files with preset tags. Supports single-tag and multi-tag classification.
-
Audio segmentation: Segment audio content and add tags to each segment.
-
Audio Recognition: Recognize and transcribe text from audio content.
-
OCR recognition results configuration
Available only when Template is Image and OCR is selected.
Default: OCR Recognition Results. Perform OCR on text within selected image areas.
Label Configuration
Enter tag names for labelers to recognize and apply. Press Enter after each tag name.
Example: To recognize cats in images, add tags "Cat", "American Shorthair", and "British Shorthair".
Specify whether labelers apply one or multiple tags to a selected object:
-
For single tag, set to Single Choice.
-
For multiple tags, set to Multiple Choice.
Example: If set to Multiple Choice, labelers can select a cat and apply both "Cat" and "American Shorthair".
NoteSingle Choice and Multiple Choice refer to tag quantity per selected object, not total selections per sample.
Enable Intelligent Labeling
Task Description
Brief description to distinguish this job.
Assign Subtask Packages
iTAG groups annotation tasks into subtask packages based on specified rules. Labelers claim subtask packages to process all annotation tasks within them.
Allocation rules:
-
Fixed size: Fixed number of tasks per subtask package.
Subtask package size based on dataset size:
-
0 to 20,000 records: 1 to 200.
-
20,000 to 100,000 records: 5 to 200.
-
100,000 to 500,000 records: 25 to 200.
-
500,000 to 1,000,000 records: 50 to 200.
-
-
Press Import Fields: Group data by selected dataset field. Data with identical field values go into the same subtask package.
-
Targeted Assignment: Assign subtask packages to specific labelers or teams.
Check Proportion
If Task Workflow includes a check step (such as Marking-Check or Marking-Inspection-Acceptance), configure the sampling ratio. Default: 100%.
User Configuration
Based on selected Task Workflow, configure labelers, checkers, accepters, or task administrators (individuals or groups). For role permissions in iTAG, see iTAG overview.
-
-
Click Create.
View task list
After creating tasks, view all current tasks on the Task Center page in iTAG. Check task status and use the Actions column to view subtask details or retrieve labeling results.

|
Area |
Task |
Description |
|
① |
Process annotation tasks |
Click Go to the iTAG Page to access the iTAG annotation page. Claim and process annotation, check, and acceptance tasks. See Process labeling tasks. |
|
② |
View labeling job status |
View status of all labeling jobs on the Task Center page. |
|
③ |
Subtask package details |
For incomplete jobs, click Subtask Details to view subtask package completion status. For incomplete subtask packages, click Transfer to transfer to another labeler or Release to release for others to claim. |
|
④ |
Retrieve labeling results View retrieval progress |
For completed jobs, click Export Labeling Result to export results. Click Obtain Data Record to view retrieval progress and results. See Export labeling result data. |
|
⑤ |
More operations |
Click |
Next steps
Claim and process annotation tasks. See Process labeling tasks.
in the Actions column to perform operations such as unpublishing or publishing jobs.