metadata
dataset_info:
features:
- name: image
dtype: image
- name: mask
dtype: image
- name: masked_image
dtype: image
splits:
- name: train
num_bytes: 35583896865.16
num_examples: 36780
download_size: 28343042907
dataset_size: 35583896865.16
configs:
- config_name: default
data_files:
- split: train
path: data/train-*