metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: pure_question
dtype: string
- name: reasoning_path
dtype: string
- name: answer
dtype: string
- name: choices
dtype: string
- name: is_correct_prompt
dtype: bool
splits:
- name: train
num_bytes: 747264917
num_examples: 702
- name: validation
num_bytes: 195658461
num_examples: 176
download_size: 289007454
dataset_size: 942923378
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*