metadata
dataset_info:
features:
- name: index
dtype: int64
- name: figure_path
dtype: string
- name: image
dtype: image
- name: caption
dtype: string
- name: question
dtype: string
- name: choice_a
dtype: string
- name: choice_b
dtype: string
- name: choice_c
dtype: string
- name: choice_d
dtype: string
- name: answer
dtype: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 12860649700
num_examples: 152603
download_size: 12794037945
dataset_size: 12860649700
configs:
- config_name: default
data_files:
- split: train
path: data/train-*