metadata
dataset_info:
features:
- name: choices
sequence: string
- name: view
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: rationale
dtype: string
- name: pil_image
dtype: image
splits:
- name: train
num_bytes: 917907502.5
num_examples: 2500
download_size: 178347542
dataset_size: 917907502.5
configs:
- config_name: default
data_files:
- split: train
path: data/train-*