metadata
dataset_info:
features:
- name: image
dtype: image
- name: mask
dtype: image
- name: masked_image
dtype: image
- name: result
dtype: image
splits:
- name: train
num_bytes: 835301492
num_examples: 2000
download_size: 969344285
dataset_size: 835301492
configs:
- config_name: default
data_files:
- split: train
path: data/train-*