metadata
dataset_info:
features:
- name: input_image
dtype: image
- name: prompt
dtype: string
- name: edited_image
dtype: image
splits:
- name: train
num_bytes: 811398038
num_examples: 16000
download_size: 810191050
dataset_size: 811398038
configs:
- config_name: default
data_files:
- split: train
path: data/train-*