Apple's Depth Pro 2 estimates sharp, metric, temporally consistent depth from any photo or video
Scobleizer · x · 2026-11-21
Apple released Depth Pro 2, a monocular depth model that produces sharp, metric, temporally consistent depth maps from ordinary photos and video instantly, with no special sensors required.
- Runs at native resolution on 2K–4K sources, with very fast inference
- Architecture: a multi-view transformer (alternating within-frame and cross-frame attention) plus a DPT decoder outputting per-frame inverse depth
- Beats prior single-frame and multi-frame methods on sharpness (Boundary F1) and temporal consistency (Jitter) averaged over Sintel and Spring
Paper, code, and models are available on Hugging Face.
More from Models
- Inkling 2 Mini Hits 49.3% on ARC-AGI-2 at $0.14 per Task — teortaxesTex · 2026-11-21
- Frontier models can decode KV-cache side channels in context; researcher shaken — davidad · 2026-11-21
- GPT-6.2 Astra tops RoboColosseum real-world robot Elo leaderboard in one day — DJiafei · 2026-11-21
- OpenAI staff coordinate vagueposts teasing a mystery model: '[Redacted] is very good at [redacted]' — Kyrannio · 2026-11-21
- LLM chain-of-thought hurts Q-matrix accuracy: 80.9% vs 87.7% expert match, study finds — alex_verem · 2026-11-21
- Mystery model Inkling 2 Mini lands on LMArena for Battle and Agent mode testing — arena · 2026-11-21