Large-scale CelebFaces Attributes (CelebA) Dataset

The CelebA (CelebFaces Attributes Dataset) contains 202,599 celebrity face images spanning 10,177 identities, each annotated with 40 binary facial attributes and 5 landmark locations, offering extensive pose variation and background clutter. This large‑scale, richly labeled collection can be used for training and evaluating computer‑vision models on tasks such as face attribute recognition, identity verification, detection, landmark localization, and face editing/synthesis.

Please cite the following if you use the data:
Liu, Z., Luo, P., Wang, X., & Tang, X. (2015). Deep Learning Face Attributes in the Wild. IEEE Conference on Computer Vision and Pattern Recognition (CVPR).