tpremoli/CelebA-attrs-160k

CelebA-128x128 CelebA with attrs at 128x128 resolution. Dataset Information The attributes are binary attributes. The dataset is already split into train/test/validation sets. This dataset has been reduced so there's 160k train samples. Citation @inproceedings{liu2015faceattributes, title = {Deep Learning Face Attributes in the Wild}, author = {Liu, Ziwei and Luo, Ping and Wang, Xiaogang and Tang, Xiaoou}, booktitle = {Proceedings of… See the full description on the dataset page: https://huggingface.co/datasets/tpremoli/CelebA-attrs-160k.