Lapis: Efficient and High-Quality Depth Estimation via Pixel-Space Diffusion with Linear Attention
kwangmoo_yi · x · 2026-09-02
Liu et al. present Lapis, a method for depth estimation using pixel-space diffusion combined with linear attention. Contrary to the expectation that generative pixel-space methods are slow, the approach achieves SOTA results with speed comparable to DA3, thanks to the effectiveness of modern one-step denoisers.
Related event: Lapis Open-Sources Efficient Pixel-Space Diffusion Depth Estimation(2 posts)→
More from Multimodal
- MiniMax H3 Generates Photoreal National Geographic-Style Landscapes of Places That Don't Exist — techhalla · 2026-09-02
- AI Clip Creation: Balancing Flow and Exaggeration — Straight-Rub1255 · 2026-09-02
- Fixing AI Comic Line Hierarchy: A Prompt Guide — tisch_eins · 2026-09-02
- Vintage AI Generative Art Piece "Open Sea" — Merzmensch · 2026-09-02
- Woven Boxes: Unique AI Visual Art Project — zzznah · 2026-09-02
- Critique of MiniMax LoRAs: Prioritize Fixing Basic Gaps Over Niche Kinks — haremlifegame · 2026-09-02