datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MUFAC
MUFAC (Machine Unlearning for Facial Age Classifier)
This repository provides a cleaned and resolution-aligned (128x128) version of the MUFAC benchmark dataset.
📘 Description
A multi-class age classification dataset featuring over 63,000 Asian facial images with annotations for age groups and personal identities.
Preprocessed facial images (128×128 resolution)
CSV files (custom_train_dataset.csv, etc.) for structured training/evaluation
Separation of… See the full description on the dataset page: https://huggingface.co/datasets/Dasool/MUFAC.SODA_images
SODA: Stereotyped Object Diagnostic Audit — Image Set
This dataset contains the 8,000 synthetic images generated for the paper
"When Cars Have Stereotypes: Auditing Demographic Bias in Objects from Text-to-Image Models"
(arXiv:2508.03483).
The images were produced with the SODA framework,
which generates everyday objects under neutral vs. demographic-conditioned prompts to measure how
demographic cues alter the visual attributes of objects produced by text-to-image models.… See the full description on the dataset page: https://huggingface.co/datasets/Dasool/SODA_images.