metadata
dataset_info:
features:
- name: images
dtype: image
- name: questions
sequence: string
- name: answers
sequence: string
- name: question_answer_pairs
list:
- name: answer
dtype: string
- name: question
dtype: string
- name: base_question
dtype: string
- name: base_answer
dtype: string
splits:
- name: train
num_bytes: 1230407840.125
num_examples: 9255
download_size: 1217151520
dataset_size: 1230407840.125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*