AbstractImageNet100 v2 (Markdown)

Summary

AbstractImageNet100 v2 is a synthetic, validation-only image dataset derived from the ImageNet100 validation split. It is a structured-Markdown prompt iteration of the AbstractImageNet100 release.

Dataset scope

• Split: validation only
• Size: 5,000 images
• Classes: 100 ImageNet classes
• Images per class: 50
• Pairing: each image is paired to an ImageNet100 validation source image

Contents and layout

The dataset contains generated PNG images and tab-separated label files, including val_labels.tsv and source_val_labels.tsv. Images correspond to the ImageNet100 validation identities used for pairing.

Intended use

The dataset is intended for research evaluation of image classification and prompt-format dataset construction on an abstract visual domain paired to ImageNet100 validation identities.

Limitations

• This is a synthetic dataset, not a replacement for natural ImageNet images.
• It contains only the documented 100-class validation release; no training split is provided by this release.
• The repository does not document a public license or redistribution terms for this dataset. Users must obtain the required permissions for the underlying source images, model, and generated artifacts before redistribution.