This dataset was generated using GRAID (Generating Reasoning questions from Analysis of Images via Discriminative artificial intelligence), a framework for creating spatial reasoning datasets from object detection annotations.
GRAID transforms raw object detection data into structured question-answer pairs that test various aspects of object… See the full description on the dataset page:
https://huggingface.co/datasets/kd7/graid-nuimages.
