metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption
dtype: string
splits:
- name: train
num_bytes: 5406889794.338
num_examples: 97527
download_size: 5412242820
dataset_size: 5406889794.338
configs:
- config_name: default
data_files:
- split: train
path: data/train-*