metadata
dataset_info:
features:
- name: text
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 7843224039.084
num_examples: 20491
download_size: 6302088359
dataset_size: 7843224039.084
configs:
- config_name: default
data_files:
- split: train
path: data/train-*