metadata
dataset_info:
features:
- name: question_id
dtype: int32
- name: question
dtype: string
- name: image
dtype: image
- name: answers
sequence: string
- name: id_image
dtype: int64
splits:
- name: test
num_bytes: 1677847253
num_examples: 5734
- name: train
num_bytes: 9746872801
num_examples: 34602
download_size: 7170106245
dataset_size: 11424720054
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
- split: train
path: data/train-*