metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 1321519958.506
num_examples: 11781
download_size: 965596255
dataset_size: 1321519958.506
configs:
- config_name: default
data_files:
- split: train
path: data/train-*