metadata
dataset_info:
features:
- name: input_image
dtype: image
- name: reference_image
dtype: image
- name: target_image
dtype: image
splits:
- name: train
num_bytes: 1549678365
num_examples: 4000
download_size: 1480171875
dataset_size: 1549678365
configs:
- config_name: default
data_files:
- split: train
path: data/train-*