metadata
dataset_info:
features:
- name: answers
dtype: string
- name: question
dtype: string
- name: questionId
dtype: string
- name: question_types
dtype: string
- name: image
struct:
- name: bytes
dtype: binary
- name: path
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 2220172440
num_examples: 56170
- name: validation
num_bytes: 555056679
num_examples: 14043
download_size: 2714020533
dataset_size: 2775229119
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*