Search for a command to run...
G2^22VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning