The only attribute of a task that cannot be changed after task creation is the task type. Other than that, all aspects of a task can be edited after creation. Read below to see how!
Editing Task Annotations
Editing a task's completed annotations is done through the dataset details page. See this part of the Datasets section to learn more about reviewing and editing existing annotations.
Editing Task Information
Editing Task Names
Project supervisors can edit a task’s name by clicking next to the name in the Tasks tab of a project. Supervisors can then enter and save the correct name in the text box that appears.
Editing Task Descriptions
Task descriptions can be edited in the right hand panel of the task details page. Click 'add description' as shown below, or hover over the existing description to reveal an edit icon.
Editing a Task's Datasets
Assigning Tasks to Datasets
To assign a task to a new dataset, click the 'add data' button on the task's datasets tab; a modal will pop up allowing you to select any datasets to which that task has not been assigned, including any datasets you choose to upload. To perform the same operation from the datasets tab of a project, see this section
Prioritizing Datasets within a Task
Prioritizing datasets within a task allows supervisors to decide the order in which datasets in that task are annotated. By default, datasets will be annotated in the order they're added to the task, and images within the dataset will be annotated alphabetically. To change the priority of a tasks datasets, simply click and drag the datasets to re-establish their order, as shown below.
Removing Datasets from Tasks
If you wish to remove a dataset from a task, or unassign that task from the dataset, navigate to the datasets tab on the task details page, click on the action menu for the dataset you wish to remove, and select 'remove from task.' The task will no longer be applied to that dataset.
Editing Task Settings
The settings tab of a task contains multiple sections to configure specific task properties. Currently supervisors can view and alter the ability of annotators to add classes from the canvas, assign tasks to specific project members, or add instructions to a task, which annotators can view from the canvas.
By default, tasks are available to any project member; however, supervisors can restrict tasks to only certain project member with task assignments. In the settings tab of a task, navigate to the 'Assignments' panel. Use the dropdown, as shown below, to select the project members who will have access to the current task. Leave all project members unchecked to restore access to everyone.
Create detailed instructions for annotators or upload them as a PDF in the instructions section of a task's settings tab. Only project members with supervisor access can edit instructions, but everyone can see them on both the image and video annotation canvases.
Enabling Rotated Bounding Boxes
For object detection tasks on image datasets, supervisors may enable rotated bounding boxes using the available checkbox in the general section under the task's settings tab.
Enabling Class Creation on the Canvas
To allow project members to add classes directly from the annotation canvas, simply check the 'Create Classes within Canvas' option in the general section of a task's settings. This will allow annotators to type in new class names to the class selector and hit enter to create classes right from the canvas.
Editing Task Availability
Assigning Task Priorities
To prioritize tasks, simply rearrange the tasks by sorting by column or clicking and dragging them, and click ‘save order’ at the bottom of the task list. Tasks at the top of the list will be prioritized when annotators hit the ‘annotate’ button of that project.
Tasks can be disabled using the toggle at the top right of the task's page, or by selecting 'disable' from the task's option menu ( ⠇) on a project's task table. Disabling a task will prevent project members from annotating it, but will not affect other tasks.
To delete a task, click on the options icon ( ⠇) and select ‘Delete.’ All annotations associated with this task will be deleted but the associated datasets will remain.