Apple Machine Learning ResearchT1·Models
78 pts
Apple introduces Depth Pro 2, a single scalable network that processes both images and video to predict metrically accurate depth maps at native resolution, simultaneously achieving boundary sharpness and temporal consistency. It requires no camera intrinsics metadata for absolute scale and produces a 2.25-megapixel depth map in 0.3 seconds on a standard GPU, with code and weights released openly.