Set up a labeling project

Set up labeling projects in the UI or using the Python SDK

To get started, create and configure a data labeling project.

Create a new project via the UI

To create a project via the Labelbox UI, follow these steps:

Step 1: Open a new project

Under the top-level Projects page, select New project. This will open up a modal for entering your project details:

Field

Description

Name

Name of the project

Description

Description of the project

Type of data to be labeled

Specify the data type (e.g., image, video, text, etc)

Where should data come from

Select one of the two methods for ingesting data into the project:

  • Datasets
  • Batch mode (preferred)

Tags (optional)

Use Project tags to enable easier organization and retrieval of the project.

When you are done entering these details, click Create to continue to the setup stage.

Project creationProject creation

Project creation

Step 2: Setup stage

To complete the project setup you will need to specify the following:

  1. Queue data for labeling (required): Depending on the ingestion mode (Batch/Dataset) you selected during project creation, you will be prompted to add a batch or dataset from Catalog (Catalog) to the project.
Setting up a batch for labelingSetting up a batch for labeling

Setting up a batch for labeling

Note that completed steps will have a strikethrough in the UI.

  1. Configure the editor: The native Editor is the recommended labeling interface and has the most robust labeling and QA tools. However, if you need to set up a custom editor, select the Custom Editor. Either way, you will be required to create or select an ontology and add labeling instructions.
Selecting the editor and ontologySelecting the editor and ontology

Selecting the editor and ontology

Once you finish these steps, the Start labeling button will be activated. However, there are two remaining optional setup steps available:

  1. Set up quality assurance tools like benchmarks and consensus (optional): This will take you to the Quality settings page where you can set up (benchmarks) and (consensus).

  2. Review and manage members of this project (optional): This will take you to the Members page within the Settings section. For information on member permissions, see Add members or groups to a project.

Create a new project via the Python SDK

import labelbox
from labelbox import LabelingFrontend, MediaType

# Create a new project
project = client.create_project(name="my-test-project",
                                description="a description",
                                media_type=MediaType.Image))

# Get the Labelbox editor (if using custom editor)
editor = next(client.get_labeling_frontends(where = LabelingFrontend.name == 'editor'))

# Get exsiting ontology (you can share an ontology across projects)
ontology = client.get_ontology("ONTOLOGY_ID")

# Get dataset
dataset = client.get_dataset("DATASET_ID")

# Setup project with editor and normalized ontology
project.setup_editor(ontology) ## Default method
project.setup(editor, ontology.normalized) ## If using custom editor

# Upload labeling instructions
project.upsert_instructions("LOCAL_FILE_PATH (PDF or HTML)")

#Attach the dataset to the project
project.datasets.connect(dataset)

#Detach the dataset from the project
project.datasets.disconnect(dataset)

Choose human workforce configuration

Workforce Boost
Many customers need to use a dedicated external data labeling team. Labelbox offers a premium data labeling service integrated tightly into the Labelbox Annotate platform.

Currently, Workforce Boost is only available to customers who have purchased the software platform. You can get a free trial anytime by contacting Support or [email protected]

Internal data labeling team
To learn how to invite, onboard, and manage your own data labeling team, see Manage members.

📘

Collaboration & rapid iteration is critical for achieving outstanding outcome with data quality.

Our very best customers run their data labeling operations through real-time collaboration with the data labeling workforce, data engineers, and ML engineers. In Labelbox, you can easily augment your internal domain experts with external data labeling service providers.

Project settings

Any time during the life of a data labeling project, you can easily update key project settings. Below are the most common things you will likely need to update:

  1. Add or remove data
  2. Update ontology
  3. Configure Model-assisted labeling (MAL)
  4. Configure QA tools

Delete a project

A project can be deleted by Admin. To delete a project, go to Danger zone under project settings.

❗️

Warning

Deleting a project will also delete all the labels that have been submitted for a project.

Set up project via the Python SDK

Python Tutorial

Github

Google Colab

Labeling projects

Open in Github

Open in Google Colab


Did this page help you?