metadata
dataset_info:
features:
- name: pos_item_ids
sequence: string
- name: pos_item_contents
sequence: string
- name: img_id
dtype: string
- name: img_path
dtype: string
- name: image_id
dtype: string
- name: question_id
dtype: string
- name: question
dtype: string
- name: answers
sequence: string
- name: gold_answer
dtype: string
- name: question_type
dtype: string
splits:
- name: train
num_bytes: 219390401
num_examples: 167369
- name: valid
num_bytes: 11341161
num_examples: 9852
- name: test
num_bytes: 4634457
num_examples: 3750
download_size: 39592933
dataset_size: 235366019
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*
- split: test
path: data/test-*