DepthART scales monocular depth to tiny models, hitting 1000 FPS on RTX A6000
kwangmoo_yi · x · 2026-07-22
Xue et al. propose DepthART, a distillation approach for scaling foundation monocular depth to tiny models.
- Uses camera-conditioned fine-tuning with a well-strategized distillation setup.
- Reported speed: 1000 FPS on RTX A6000 and 200 FPS on Jetson Orin NX.
- The post frames it as a practical way to bring strong monocular depth performance to lightweight deployment targets.
Related event: DepthART Distills Monocular Depth to Tiny Models(2 posts)→
More from Research
- Stanford Team Introduces Gigatoken, the World's Fastest Tokenizer — StanfordAILab · 2026-07-22
- Tabul AI launches Metal TreeSHAP to speed up Shapley values on Apple silicon — Scobleizer · 2026-07-22
- Reddit points to OpenAI’s ChatGPT Ads page — EcstaticAsparagus509 · 2026-07-22
- Open-source runtime lets each repo define its own AI code reviewer — ibabufrik · 2026-07-22
- DeepSWE: A New Benchmark for Evaluating AI Coding Agents on Real GitHub Issues — pmz · 2026-07-22
- A Rust space-economy sim runs hundreds of autonomous ships, built with Claude — kalcode · 2026-07-22