Models / meta / DPT DINOv2 Giant NYU

DPT DINOv2 Giant NYU

JSON →
metavision

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

imagevision
Specs
Context & limits
context window
max output
Lifecycle
Dates
released2024-04-01
Resources