✨
AI Summary
- DINOv3 is a self-supervised learning model for robust visual representations.
- Learns from unlabeled data, reducing the need for human annotation.
- Introduces Gram anchoring to maintain high-quality dense features.