Randomized-MLP Regularization Improves Domain Adaptation and Interpretability in DINOv2

Open in new window