Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 36 Next »

Manage datasets containing the project input and output data.

Use the Datasets tab to manage data that you feed into your project.

You must have the Create/Edit Batches privilege to be able to work with this feature.

The Datasets page organizes the datasets that you added to the project into the following tabs. Click on each for more information.

  • Active: Datasets on which work is currently progressing.

  • Pending: Datasets that have been imported, but on which work is yet to start.

  • Completed: Datasets on which work is completed.

  • Archived: Datasets on which work is completed and which have been archived for long-term storage. You can archive a pending batch, but you cannot archive an active batch. To archive active batches, you must first cancel the active batch and then archive it.

  • Paused: Datasets on which work started but was paused.

  • Cancelled: Datasets that were imported and then cancelled.

Adding Dataset Batches

Any newly-added batch automatically appears in the Pending tab.

To add a dataset to your project:

  1. Click the Add Batch button in the top-right corner of the Datasets tab. The Add Batch modal appears.

  2. Enter the name of the batch in the Batch Name field.

  3. Click to select the Priority of the batch (optional). The higher the priority value of a batch, the earlier it gets routed to labelers for processing.

  4. Enter a Message that you want to appear when labelers start work on this batch (optional).

  5. Enter the Email ID that must be notified when the tasks are imported and when work on the batch is completed (optional).

  6. Click the ETA field and select the date by which work on this batch must be completed (optional). This will be saved as the target date of this batch.

  7. Click to specify whether this batch is a Golden Batch (optional).

  8. Click Submit to add the dataset batch to the project.

Export Project Output

To export the output created for a project:

  1. Click the Ellipsis icon at the top-right corner of the Datasets table and select Export Project Output.

  2. The Export Project modal appears.

3. All is selected by default for Select Batch.

4. Select CSV from the list for Select format.

5. Select the field(s) you want to export for Select fields. By default, all the fields will be exported.

6. (Optional) Select the values for Start date and End date.

7. Select the Completed tasks only checkbox to download the data associated with only the completed tasks in your project. If not selected, all the tasks will be exported.

8. Click Download to trigger the export. Once the export completes, the modal notifies you and offers a link to download the exported content.

9. Click Download File to download the exported output.

10. Click Close to return to the Dataset tab.

Configure Cloud Storage

To configure cloud storage options for your project:

  1. Click the Ellipsis icon in the top-right corner of the Datasets table and select Configure Cloud Storage. The Cloud Storage Configuration modal appears.


    The Cloud Storage Configuration modal enables you to configure the connection details that are required for you to connect with specific cloud service providers. Taskmonk enables you to store your projects with the following service providers:

    • Amazon S3

    • Microsoft Azure

    • Google Cloud

  2. Click to select the service provider that you want to use to upload your project data.

    Depending on your choice from these options, the fields displayed change.

Configuring Amazon S3 Cloud Storage in Taskmonk

  1. Enter the Bucket Name in which you want to store your project data.

  2. Enter the Region where you created the bucket. For details on the list of regions available in Amazon S3, see Amazon documentation.

  3. Enter the Access Key and Secret Key associated with your account.

  4. Click Submit to save your changes and return to the Datasets tab.

Configuring Microsoft Azure Cloud Storage in Taskmonk

  1. Enter the Account Name and Account Key provided by your service provider in the fields provided.

  2. Click Submit to save your changes and return to the Datasets tab.

Configuring Google Cloud Storage in Taskmonk

  1. Enter the Bucket Name and Auth String in the fields provided.

  2. Click Submit to save your changes and return to the Datasets tab.

Dataset Status Tabs

 Archived Datasets
 Paused Datasets

Paused Datasets

The Paused tab displays all the datasets for which work is paused and enables you to perform the following tasks:

Key Batch Tasks

 Edit / Update a Batch

Edit / Update a Batch

To edit or update a batch:

  1. Click the Edit icon in the row representing the batch. This enables all the editable fields associated with the batch.

    Note: All steps in this section are optional.

  2. Update the Name of the batch.

  3. Click to update the Priority of the batch. The higher the priority value of a batch, the earlier it gets routed to labelers for processing.

  4. Click the ETA field and select the date by which work on this batch must be completed. This will be saved as the target date of this batch.

  5. Click the Update icon to save your changes and update the batch.

 Delete a Batch

Delete a Batch

To delete a batch, click the Delete icon in the row representing the batch and click Yes in the confirmation modal that appears.

 Assign a Batch to Specific Users

Assign a Batch to Specific Users

To assign a batch to specific users, click the Privileged button in the row representing the batch. When you enable this option, the Add Users option is added to the list of Batch Config tasks associated with the batch. For information on working with this option, see Adding Users to a Batch.

 Import Tasks into a Batch

Import Tasks into a Batch

To import tasks into a batch:

  1. Click the Import button in the row representing the batch. The Import Task modal appears.

  2. Select the radio button corresponding to the file type that you want to use to import tasks.

  3. Click Choose Files and select the file that contains the tasks that you want to import.

    Note:

    1. The column names in your import file must match your project’s input field names. If you are unsure about the structure of the input file, click the Download Template File link below the radio buttons. Once you download this template, you can update it using the data that you want to import. You must, however, ensure that your data is organized to match the field labels in the template file.

    2. For media file imports, the allowed file type depends on the field type set in the input fields. For example, video field types allow you to upload only MP4 and AVI files.

  4. Click Import. This triggers the import, and the modal informs you once the import is successful.

  5. Click Close to return to the Datasets tab.

 Export Tasks from a Batch

Export Tasks from a Batch

You can also download the tasks that you have imported into Taskmonk, regardless of whether they are completed or not.

This does not delete the tasks in Taskmonk.

To download an imported task:

  1. Click the Export button in the row representing the batch. The Import Task modal appears.

  2. Select the radio button corresponding to the format in which you want to download the exported tasks.

  3. Enter the Start Date and End Date to specify that you want to download only those items in the dataset on which work was done between the specified dates (optional).

  4. Click Select Columns to choose the columns whose data you want to download. If you want to download the data in all the columns, select the Select All Fields check box.

  5. Select Completed Tasks Only to download only those tasks that are complete.

  6. Click Download to export the file.

  7. Once the file is exported, click Download File to save the exported file to your machine.

  8. Click Close to exit the modal and return to the Datasets tab.

 View Batch Tasks

View Batch Tasks

To view the tasks in a batch:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click View Batch Task. The Batch Tasks modal appears.

  2. Use the search fields provided below each column to search for data that matches specific criteria. For example, in the screenshot above, the Product_Title column is filtered using the string Fountain. You can similarly use multiple filters to narrow down the data that appears.

  3. Use the Back and Forward icons to move between pages as required.

 Upload Batch Documents

Upload Batch Documents

To upload a document specific to a batch:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Upload Batch Documents. The Upload Document modal appears.

  2. Enter the Document Name in the field provided.

  3. Click Browse and choose the file that you want to upload.

  4. Click Upload Document to add your document to the batch and return to the Datasets tab.

 View Batch Documents

View Batch Documents

To view the documents associated with a batch:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click View Batch Documents. The View Batch Documents modal appears.

  2. Click the View Item icon to view the document in a separate modal. You can even download the uploaded file by clicking the Download Item icon.

  3. To delete an uploaded item, click the Delete Item icon.

  4. Click Close once you are done.

 Import Output

Import Output

If you have already done some annotation, you may want to include it into a batch by importing it. To import project output into a batch:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Import Output. The Import Output modal appears.

  2. Click to select the format in which you want to import the output. You can choose between Excel and CSV.

  3. Click Choose Files to select the files that you want to import.

  4. Click the Select Level field and choose the output levels that you want to import.

  5. Click Import. The data in the files you selected will now be imported into the task execution levels of your choice. When team members working on the next level log in, they will find that the batch is ready for their work, as the work associated with the previous level is already complete.

 Export to Customer

Export to Customer

To export the output of a batch to a customer:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Export to Customer. The Export to Customer modal appears.

  2. Click to select the format in which you want to export the output. Choose from the available options, CSV and Excel.

  3. Use the Start Date and End Date fields to select the start and end dates between which the data that you want to export was created.

  4. Click the Select Column field to select the fields that contain the data that you want to export. If you want to export the data in all the fields, choose Select all fields.

    This will include only those fields for which Customer Visible is marked as True in the Task Design > Field Settings page.

  5. Select the Completed Tasks Only check box to indicate that you want to export only completed tasks.

  6. Click Download to create the exported file.

  7. Once the file is ready, the Download File link appears. Click this link to download the exported file.

  8. Click Close to return to the Datasets page.

 Change State

Change State

To change the status of a batch:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Change State. The Change Batch State modal appears.


    The Change Batch State modal displays the current status of the batch and enables you to change it as required.

  2. Click the Change State to and choose the new state that you want to assign to the batch.

Depending on the initial state of the batch, the target states change. For details on the options available to you in each state, see the table below:

Current State

Target State Options

Comments

Active

Pause, Canceled

When an active batch is paused, tasks from the batch are no longer passed to team members; but the progress made is saved.

Pending

Pause, Canceled

When a pending batch is paused, the batch is taken off the queue and does not automatically become active until its status is changed to Resume.

When a batch is canceled, it is taken off the queue and can only be moved to Pending. The progress made, however, is saved.

Completed

Archived

Archived

Pending

When an archived batch is moved to Pending status, it is reintroduced into the queue. The progress made before archival is retained.

Paused

Resume, Canceled

Canceled

Pending

When a canceled batch is moved to Pending, the batch rejoins the queue. All the progress made before cancelation is retained.

 Add Users to a Batch

Add Users to a Batch

To add specific users to a batch:

  1. Enable the Privileged toggle in the batch where you want to add users.

  2. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Add Users. The User Allocation to Privileged Batch modal appears.

    The User Allocation to Privileged Batch modal enables you to add specific users to a batch. The users listed here must have been previously added to the project.

  3. Click the Assign Users to Batch and select the users whom you want to assign to the batch.

  4. Click Save.

 Specify Output Directories

Specify Output Directories

To specify directories in which the output from a batch must be written:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Output Directory. The Output Directory modal appears.

  2. Enter the Folder Names into which you want to write the output and specify the Percentage of data that must be written into each folder. Click Add Folder to create new folders.

  3. Click Save to enforce your configurations and return to the Datasets tab.

 Add a Task Submission Confirmation Message

Add a Task Submission Confirmation Message

A task submission message is an optional prompt that appears when a labeler submits a task.

To add a confirmation message to tasks associated with a batch:

  1. Click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Task Submission > Confirmation Message. The Confirmation Message modal appears.

  2. By default, this option is disabled. To enable confirmation messages, select Enabled in the Confirmation Message field.

  3. Enter the confirmation message that you want to be displayed when the labeler submits a task in the field provided.

  4. Click Submit to save the message and return to the Datasets tab.

 View as Gallery

View as Gallery

If you are working on an image-related project, the Gallery View enables you to view the images that you imported in a batch, and perform specific tasks related to the images in the batch.

To access the gallery view, click the Batch Config Settings icon in the row representing the batch and, from the menu options that appear, click Gallery View. The Gallery View modal appears.

The Gallery View modal offers controls that enable you to perform the following tasks:

  1. Search for specific images

  2. View labeling input and output

  3. Revalidate completed labeling

  4. Download the dataset

  5. Regenerate output for completed tasks

Search for Images

To search for an image in the Gallery, enter any part of the name of the image in the Search field. The modal displays a filtered list of images in real time.

View labeling input and output

The Gallery View modal displays both input images and the output that labelers created for them.

Viewing Labeling Input

The input images appear by default when the Gallery View modal appears. To redisplay the input view if you are viewing the output, click the View and select Input.

Viewing Labeling Output

To view the output view and related details:

  1. Click the View and select Output. The modal now displays the images that have been labeled while displaying placeholder images for those that haven’t been labeled yet.

  2. To view the labeling details associated with an image, click it. The image opens in a new modal named after the image, and you can see both the input image, output labels or segmentation, and the overlapped visualization view.

  3. To download the JSON file generated by the labeling, click Download JSON.

  4. Click Regenerate Output to regenerate the labeling output. This is especially useful if you decide to change the output format of the labeling data. For example, if you changed the output format from COCO to Pascal VOC, or any other custom format, click this button to regenerate the labeling data in the new preferred format.

  5. Click OK to close the modal and return to the Gallery View modal.

Download the dataset

To download the dataset, complete with the labels, visualizations, and so on, click the Download Dataset button. All the data associated with the batch downloads on to your computer as a ZIP file.

Regenerate output for completed tasks

If you change the output format settings associated with your project mid-way into a batch, you may want to regenerate the output to match the new format requirements. To do so, click Regenerate Output for Completed Tasks and click Yes in the confirmation modal that appears.

  • No labels