metadata
dataset_info:
features:
- name: text
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 1591414771.625
num_examples: 3817395
download_size: 1401834720
dataset_size: 1591414771.625
configs:
- config_name: default
data_files:
- split: train
path: data/train-*