GraphNav: Benchmarking Spatial Cognitive Graph Reasoning in Vision-Language Models
Overview
GraphNav is a controlled benchmark designed to evaluate spatial cognitive-graph reasoning in vision-language models (VLMs) # GraphNav: Benchmarking Spatial Cognitive Graph Reasoning in Vision-Language Models
Overview
GraphNav is a controlled benchmark designed to evaluate spatial cognitive-graph reasoning in vision-language models (VLMs) by isolating high-level… See the full description on the dataset page: https://huggingface.co/datasets/graphnav-anonymous/GraphNav.
