DPT DINOv2 Giant NYU

JSON →
meta vision
image

A giant Dense Prediction Transformer (DPT) model using DINOv2 backbone, fine-tuned on NYU Depth v2 for monocular depth estimation.

vision
releasedApr 2024