metadata
dataset_info:
features:
- name: image
dtype: image
splits:
- name: test
num_bytes: 274664364.23
num_examples: 10130
- name: train
num_bytes: 5216078696.499
num_examples: 192469
download_size: 0
dataset_size: 5490743060.729
Dataset Card for "CelebA-faces-cropped-128"
Just a 128px version of the CelebA-faces dataset, which I've cropped to the face regions using dlib. Processing notebook: https://colab.research.google.com/drive/1-P5mKb5VEQrzCmpx5QWomlq0-WNXaSxn?usp=sharing