metadata
dataset_info:
features:
- name: dataset
dtype: string
- name: scene_name
dtype: string
- name: question
dtype: string
- name: ground_truth
dtype: string
- name: question_type
dtype: string
splits:
- name: train
num_bytes: 810674
num_examples: 3519
- name: test
num_bytes: 810674
num_examples: 3519
download_size: 485964
dataset_size: 1621348
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*