Robbyant/lingbot-map

Health v2 · maintenance only

Flags: no_releases young

Adoption not part of the score

16705 stars · 1864 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

LingBot-Map is a feed-forward 3D foundation model that reconstructs scenes from streaming image data using a Geometric Context Transformer. It performs real-time streaming 3D reconstruction at roughly 20 FPS over sequences exceeding 10,000 frames, with paged KV cache attention for long-sequence stability.

Use cases

• reconstruct 3D scenes from streaming video
• run real-time 3D reconstruction on long sequences
• estimate camera trajectory and correct drift in video
• generate dense geometry from monocular or multi-view streams
• evaluate 3D reconstruction on KITTI or Oxford Spires benchmarks
• build mapping for robotics or autonomous navigation

When to choose

• you need feed-forward streaming 3D reconstruction rather than slow iterative optimization
• you process very long video sequences (thousands of frames) and need stable memory usage
• you want state-of-the-art reconstruction quality with real-time inference on GPU
• you need a permissively licensed (Apache-2.0) 3D foundation model with pretrained weights

When to avoid

• you need lightweight CPU-only inference, since the model targets GPU acceleration
• you need a full production SLAM system with loop closure and sensor fusion beyond visual reconstruction
• you need a mature stable release rather than a fast-moving research codebase

Facets

library · maturity active

machine-learning deep-learning computer-vision image-processing simulation computer-vision artificial-intelligence deep-learning robotics autonomous-vehicles python 3d-reconstruction streaming-inference foundation-model slam transformer depth-estimation point-cloud kv-cache gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Robbyant/lingbot-map")