metadata
dataset_info:
features:
- name: image
dtype: image
- name: true_caption
dtype: string
- name: false_caption
dtype: string
splits:
- name: train
num_bytes: 1410724064
num_examples: 800
download_size: 1410108137
dataset_size: 1410724064
configs:
- config_name: default
data_files:
- split: train
path: data/train-*