metadata
dataset_info:
features:
- name: image
dtype: image
- name: markdown
dtype: string
- name: language
dtype: string
- name: source
dtype: string
- name: doc_id
dtype: string
- name: page_num
dtype: int32
splits:
- name: train
num_bytes: 550060499
num_examples: 3813
- name: test
num_bytes: 68519001
num_examples: 425
download_size: 683792455
dataset_size: 618579500
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*