Appearance

Getting Started with Labelo ​

Welcome to Labelo, an open-source data labeling tool designed for seamless project management and collaboration. This guide will help you get started with setting up your Labelo environment and launching your first data labeling project.

1. Install Labelo ​

You can install Labelo on your local machine or server. Labelo provides flexibility to run it either using Docker or installing it directly using pip.

Install using pip: ​

bash
pip install labelo
labelo

Install using Docker: ​

bash
docker pull cybrosys/labelo:latest
docker run -p 8080:8080 -v $(pwd)/mydata:/labelo/data cybrosys/labelo:latest

After installation, Labelo will be accessible via http://localhost:8080.

2. Create a New Project ​

Once Labelo is up and running:

  • Log in and navigate to the Project page.
  • Click Create Project and enter a project name and description.
  • Select a workspace for the project.

3. Import Data ​

  • Go to your project and click on Import Data.
  • You can upload various data formats (e.g., images, text files, audio, video) by dragging and dropping or selecting files from your system.

Labelo supports data types such as:

  • Text: .txt
  • Audio: .wav, .mp3, .flac, .m4a, .ogg
  • Video: .mp4, .mpeg4, .webp, .webm
  • Images: .jpg, .jpeg, .png, .gif, .bmp, .svg, .webp
  • HTML: .html, .htm, .xml
  • Time Series: .csv, .tsv
  • Common Formats: .csv, .tsv, .json

4. Configure the Labeling Interface ​

Customize the labeling interface to match your project’s needs:

  • Set up tools for labeling, such as bounding boxes for images or text classification for documents.
  • Define label options, making sure all categories are accurately represented for labeling.

5. Start Labeling ​

Once your project and data are set up:

  • Use Labelo’s Editor to start labeling your data.
  • Apply the appropriate labels based on your configuration. The intuitive interface makes it easy to add comments, tags, and annotations to your data.

6. Review Annotations ​

After labeling:

  • Review the annotations to ensure accuracy.
  • Labelo allows you to approve or reject annotations and add comments for feedback. The built-in review system helps maintain quality across your data annotations.

7. Export Labeled Data ​

When the project is complete:

  • Export your labeled data in various formats such as JSON, CSV, COCO, or YOLO, depending on your machine learning workflow.
  • Labelo’s flexible export options ensure compatibility with most popular ML frameworks.

8. Manage Teams and Workspaces ​

Labelo supports multi-user collaboration:

  • Invite team members and assign them roles (e.g., Annotator, Reviewer, Manager).
  • Organize your projects within Workspaces, making it easy to manage tasks and separate different workflows.

Tips for Success:

  • Customization: Adjust Labelo’s interface to suit your project’s unique needs.
  • Integration: Labelo can be integrated into your existing ML pipeline.
  • Community Support: Being open-source, Labelo has an active community that you can reach out to for help or suggestions.