Depth Anything 3 (DA3-LARGE) is a 0.35B parameter foundation model developed by ByteDance for unified depth and pose estimation. It uses a novel depth-ray representation and supports both relative depth prediction and camera pose estimation from single or multiple views. The model ships with a Python package, CLI tools, and is available as open weights on Hugging Face.
In the Other AI space, DA3 LARGE takes a focused approach. It focuses on estimating accurate depth maps and camera poses from one or more images without specialized hardware. DA3 LARGE is an open-source project aimed at developers. The project is open source (Apache-2.0). It runs on the web, the command line, and API.
It is developed by ByteDance (China), and it first shipped in 2025. Development happens publicly on GitHub with 5.9k stars and 2 commits in the last 90 days. Among its 4 catalogued features are monocular depth estimation, multi-view geometry, and pose estimation.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do