A novel VQA benchmark for VLMs that evaluates models on their embodiment-specific understanding of navigation across challenging real-world scenarios.
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
DeFM: Learning Foundation Representations from Depth for Robotics
NaviTrace: Evaluating Embodied Navigation of Vision-Language Models