DINOv2
Label-free features that beat supervised pipelines
Latest: DINOv2 with registers (2023-24); superseded by DINOv3
A family of self-supervised vision backbones (up to a 1.1B-parameter ViT-g) trained on the curated 142M-image LVD-142M dataset, producing all-purpose features that beat supervised and CLIP-style models on classification, depth, and segmentation without fine-tuning. Initially research-licensed in April 2023, relicensed Apache 2.0 later that year.
Why it matters
DINOv2 proved frozen, label-free features could beat supervised and CLIP-style pipelines on classification, depth and segmentation without fine-tuning. Relicensed Apache 2.0, it spread from medical imaging to forest-canopy mapping with the World Resources Institute and Mars-robotics work at NASA JPL, making it the practical default vision backbone of 2023-25.
Facts
- It learned to see — matching parts across totally different objects — without ever reading a single label.
- DINOv2 features were adopted for medical imaging (histology, endoscopy) and depth estimation; its RGB-coded PCA-of-features videos went viral among researchers as a way to 'see' what a network sees.
- Proved frozen, label-free features could beat supervised and text-supervised pipelines; 26 authors (Oquab et al.).
- Used with the World Resources Institute to map forests and estimate tree canopy height from satellite imagery; NASA JPL adopted DINOv2 features for Mars exploration robots.
Try it yourself
Official DINOv2 demo ↗ dinov2-base on Hugging Face ↗ Run the segmentation notebook in Colab ↗
Lineage
Sources
arXiv ↗GitHub · dinov2 ↗Meta AI blog ↗