metadata
dataset_info:
features:
- name: filename
dtype: string
- name: page_num
dtype: int64
- name: qa_pair
struct:
- name: question
dtype: string
- name: answer
dtype: string
- name: image
dtype: image
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 863796572.404
num_examples: 1059
download_size: 861680695
dataset_size: 863796572.404
configs:
- config_name: default
data_files:
- split: train
path: data/train-*