metadata
dataset_info:
features:
- name: image
dtype: image
- name: text
dtype: string
splits:
- name: train
num_bytes: 21006763603
num_examples: 90000
download_size: 13660207392
dataset_size: 21006763603
configs:
- config_name: default
data_files:
- split: train
path: data/train-*