GroundingSuite: Measuring Complex Multi-Granular Pixel Grounding
Paper β’ 2503.10596 β’ Published β’ 18
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
GSEval is a meticulously curated evaluation benchmark consisting of 3,800 images, designed to assess the performance of pixel-level and bounding box-level grounding models. It evaluates how well AI systems can understand and localize objects or regions in images based on natural language descriptions.
git lfs install
git clone https://huggingface.co/datasets/hustvl/GSEval