A Cheat Sheet for Spatial Reasoning: GraFT Gives Vision-Language Models a 3D Map Instead of Retraining Them

A Cheat Sheet for Spatial Reasoning: GraFT Gives Vision-Language Models a 3D Map Instead of Retraining Them | BedrockNews