Renaissance Labs

Atlas / Europe

Independent company · Zurich, Switzerland

Lightly

Lightly is a Zurich company, born from ETH research, that open-sources self-supervised learning and vision training libraries.

computer visionself-supervised learning

Zurich, Switzerland

Type
Independent company
Based in
Zurich, Switzerland
Featured projects
8

Update Lightly is a commercial startup. Some of its research code and models are open source.

About

Lightly was founded by Matthias Heller and Igor Susmelj as a data-centric AI company based on ETH Zurich research. It raised a 3 million dollar seed round in May 2022 and reports 10 million dollars raised in total.

Its open-source lightly library for self-supervised learning dates from October 2020. It later added the labelformat converter, LightlyTrain for vision model pretraining and distillation, and LightlyStudio, alongside a commercial platform.

Mission and ambitions

To help teams build better vision models by improving the data that models are trained on.

Known for

  1. 01

    Lightly self-supervised learning library

  2. 02

    LightlyTrain vision pretraining

  3. 03

    ETH Zurich spin-off

Milestones

  1. 2020

    lightly library

    The open-source self-supervised learning library was published on GitHub.

  2. 2022

    Seed round

    Lightly secured 3 million dollars led by Wingman Ventures.

  3. 2023

    labelformat

    A converter between computer vision label formats was released.

  4. 2025

    LightlyTrain

    A training toolkit for YOLO, ViTs and RT-DETR was released.

  5. 2025

    LightlyStudio

    A unified data platform for multimodal ML was published.

Open source

GitHubgithub.com/lightly-ai
  • A python library for self-supervised learning and latent world models on images. (3,832 stars)

  • All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation. (1,741 stars)

  • LightlyStudio - The Unified Data Platform for Multimodal ML (911 stars)

  • A tool for converting computer vision label formats. (91 stars)

Hugging Facehuggingface.co/lightly-ai

Notable releases

Openness

Open-source libraries with a commercial platform alongside.

People

Listed from the lab's own pages and public profiles.

Contact and community

Contact through the company website.

Sources

Compiled from the lab's public pages, Hugging Face and GitHub. Last checked 2026-11-17. Something wrong or out of date? Tell us on Discord.

Labs nearby