metadata
dataset_info:
features:
- name: image
dtype: image
- name: id
dtype: string
- name: query
struct:
- name: question
dtype: string
- name: answers
sequence: string
- name: words
sequence: string
- name: bounding_boxes
sequence:
sequence: int64
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 124487484.92307693
num_examples: 180
- name: test
num_bytes: 13831942.76923077
num_examples: 20
download_size: 148303453
dataset_size: 138319427.6923077
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*