metadata
dataset_info:
features:
- name: image
dtype: image
- name: qa
list:
- name: question
dtype: string
- name: answer
dtype: string
- name: fullAnswer
dtype: string
splits:
- name: train_balanced
num_bytes: 10172551012.98
num_examples: 72140
- name: val_balanced
num_bytes: 1429170102.54
num_examples: 10234
download_size: 11650910411
dataset_size: 11601721115.52
configs:
- config_name: default
data_files:
- split: train_balanced
path: data/train_balanced-*
- split: val_balanced
path: data/val_balanced-*