LabelLab

LabelLab is an image analysis and classification platform for Machine Learning researchers. Originally built to classify images of elephants, it has been developed over the years and is now a full-fledged ML utility tool that is able to classify all kinds of images and carry out a number of ML-related tasks in a project-based manner while teaming up with other researchers.

Features

• The user can create a project and add other users as collaborators.
• Inside a project, multiple actions can be performed such as -
  • Uploading images in batches
  • Creating new labels
  • Labelling the uploaded images
  • Creating and training different kinds of models
  • Using the image path tracking feature on labeled images
• Each collaborator can be assigned one or more roles such as images, labels, models, admin, etc which will allow that collaborator to perform tasks specific to that role.
• On basis of their roles, the collaborators are divided into various teams.
• Each team has a chat system to discuss the assigned tasks and progress on the project.
• The project also has an activity tracking system recording all progress in it, which can be filtered by team, project member, category, and entity.
• Besides that, the user can also upload images outside a project and run classifications against a trained model.

Apache 2.0 Licence

See the LICENSE file for details.

Deployment

Host your own LabelLab server for FREE with One-Click Deploy

Documentation

To learn more about the project, check out the Project Wiki

User Guide

Please go though our Code of Conduct first.

Follow these instructions to setup the project in your local system.

• Clone the repository.

git clone https://github.com/scorelab/Labellab.git

• Follow the instructions provided below for setting up each section of the project

  • Backend Setup

  • Frontend Setup

  • Mobile Setup

NOTE: The Node.js API in labellab-server directory is now deprecated and is no longer being used in the project. Instead, use the Flask API in labellab-flask directory.

Contact

• Join the LabelLab Gitter Channel for any queries or discussions regarding the project.
• For learning more about SCoRE Lab and its various other projects, check out our Website