Overview

DINOv2 is a self-supervised method for training computer vision models developed by Meta Research and released in April 2023. Because DINOv2 is self-supervised, the input data does not require labels, which means models based on this architecture can learn richer information about the contents of an image.

Performance

Use This Model

Label Data Automatically with DINOv2

You can automatically label a dataset using DINOv2 with help from Autodistill, an open source package for training computer vision models. You can label a folder of images automatically with only a few lines of code. Below, see our tutorials that demonstrate how to use DINOv2 to train a computer vision model.

No items found.

DINOv2

Overview

Performance

Use This Model

Label Data Automatically with DINOv2

Deploy to Production

Data

Training

Utilities

Compare to Related Models

Convert Annotation Format

Architecture

Frameworks

Annotation Format

Explore This Model on Roboflow

Join over 250,000 developers managing computer vision data on Roboflow.

Try It for Free