This dataset comprises images and annotations from the original Open Images Dataset V7.
Out of the 9M images, a subset of 1.9M images has been annotated with automatic methods (Image-text-to-text models).
This dataset comprises all 1.9M images with bounding boxes annotations
from the Open Images V7 project.
The annotations… See the full description on the dataset page:
https://huggingface.co/datasets/Fhrozen/openimages-narratives-v2.
