lightly/worker repository overview

⁠What is Lightly?

The Lightly solution⁠ is an intelligent computer vision system designed to process raw, unlabeled image data, select the most informative samples for labeling, and mitigate dataset bias. Lightly scales to big datasets with millions of images or thousands of videos. Processing these large datasets requires a special architecture. This repository hosts the image of the Lightly Worker.

⁠Lightly Worker

The Lightly Worker is a Docker container designed to process large datasets. You host it yourself on a machine of your choice. The Lightly Worker processes jobs from a job queue and stores the outputs back to your cloud or local storage. Your sensitive data never leaves your cloud storage.

⁠Getting Started

The best way to getting started with Lightly is to follow our getting started⁠ guide.

⁠Documentation

The Lightly Worker features an extensive and advanced documentation⁠.

⁠Features:

• Select from more than 10 million samples within a few hours!
• Select by diversity, weights, threshold or similarity
• Active Learning pipeline built in
• Runs directly with videos without prior extraction of the frames!
• Enterprise ready
• Secure
• Built-in pretagging models
  • Use this feature to pre-label your dataset or to only select images which contain certain objects
  • Supported object categories are: bicycle, bus, car, motorcycle, person, train, truck
  • Supports your own predictions
• Support for custom metadata
• Automated reporting of the datasets for each run with
• PDF report with histograms, plots, statistics, and much more …
• Multithreaded
• SIMD instructions
• And much more⁠

⁠Changelog

Stay informed about our newest features and updates by checking out our changelog⁠.

⁠Hardware Requirements

Please refer to our hardware requirements⁠ to choose the right hardware.

⁠Security

Please refer to our security document⁠ to understand how Lightly does not need or want to see your sensitive data.

⁠Support And Contact

You can reach out to us anytime via our Discord⁠ channel or contact us⁠ directly.