This dataset is composed of scores of images taken from English Wikipedia and Wikimedia Commons. The scores are the outputs of the models
manual curation of images in commons that are either explicit or likely to be misflagged as explicit
taking prominent images from the top ~300k English Wikipedia article… See the full description on the dataset page:
https://huggingface.co/datasets/derenrich/enwiki-image-content-moderation.