metadata
dataset_info:
features:
- name: original_caption
dtype: string
- name: edit_prompt
dtype: string
- name: edited_caption
dtype: string
splits:
- name: train
num_bytes: 1760186
num_examples: 8831
download_size: 492487
dataset_size: 1760186
configs:
- config_name: default
data_files:
- split: train
path: data/train-*