LingBot-Depth
- AI model
- Perception
A depth model that turns incomplete, noisy depth sensor readings into complete, metric-accurate 3D measurements, aimed at robot learning and 3D vision.
Made by Robbyant.
Source on GitHub: github.com/Robbyant/lingbot-depth (opens in a new tab). Facts read at commit f3a237e (opens in a new tab).
Licenses
- Code
- Apache-2.0 (opens in a new tab)
Read by Robotheca from LICENSE (opens in a new tab) (Apache-2.0) at commit f3a237e (opens in a new tab), on 16 Oct 2026.
- Weights
- Apache-2.0 (opens in a new tab)
Read on huggingface.co/robbyant/lingbot-depth-pretrain-vitl-14 (opens in a new tab).
- Commercial use
- Yes
What it does
- Refines raw sensor depth into cleaner, more complete measurements by jointly aligning RGB appearance with depth geometry.
- Fills missing depth regions with metric accuracy, including a variant optimized for sparse depth completion.
- Supports scene reconstruction, 4D point tracking of dynamic targets, and robotic grasping demos.
- Provides pretrained checkpoints on Hugging Face and ModelScope, loadable through a Python inference API.
- Trained on a curated RGB-D dataset of about 3 million real and simulated samples, released publicly.
Written by Robotheca from the project's README (opens in a new tab) at commit f3a237e.
Getting started
- Difficulty
- Intermediate
- Hardware it needs, as the project states it
- Python 3.9 or newer and PyTorch 2.0 or newer; a CUDA-capable GPU is recommended, with CPU as a fallback.
- How to start
- The project's website (opens in a new tab)
Maintenance
- Status
- Maintained: a commit in the last 12 months
- Newest commit
- 8 Jul 2026 on main
- Latest release
- None published on GitHub
Checked on GitHub on 16 Oct 2026.
Links
Robots that can run it
No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.
Robotheca lists AI models as a reference: nothing is hosted or run here, and every fact links to where it was read.