metadata
dataset_info:
features:
- name: image
dtype: image
- name: text
dtype: string
- name: language
dtype: string
splits:
- name: train
num_bytes: 11809321371
num_examples: 25380
- name: validation
num_bytes: 860113341
num_examples: 1119
- name: test
num_bytes: 755709817
num_examples: 1010
download_size: 22387477406
dataset_size: 13425144529
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*