metadata
dataset_info:
features:
- name: input_image
dtype: image
- name: target_mask
dtype: image
- name: reference
dtype: image
splits:
- name: train
num_bytes: 731210737.872
num_examples: 3928
download_size: 449008421
dataset_size: 731210737.872
configs:
- config_name: default
data_files:
- split: train
path: data/train-*