metadata
dataset_info:
features:
- name: image
dtype: image
- name: query
dtype: string
- name: answer
dtype: string
- name: pdf_name
dtype: string
- name: orig_caption
dtype: string
splits:
- name: train
num_bytes: 333764
num_examples: 5
download_size: 336640
dataset_size: 333764
configs:
- config_name: default
data_files:
- split: train
path: data/train-*