metadata
dataset_info:
features:
- name: image
dtype: image
- name: text
dtype: string
splits:
- name: train
num_bytes: 25171082933
num_examples: 21535590
- name: validation
num_bytes: 6856332
num_examples: 6000
download_size: 21613339213
dataset_size: 25177939265
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*