metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption
dtype: string
splits:
- name: train
num_bytes: 4977602607.549
num_examples: 47891
download_size: 6141914429
dataset_size: 4977602607.549
configs:
- config_name: default
data_files:
- split: train
path: data/train-*