metadata
dataset_info:
features:
- name: image
dtype: image
- name: doc
dtype: int64
- name: folder
dtype: string
- name: label
dtype: int64
- name: input_ids
dtype: string
- name: bbox
dtype: string
- name: attention_mask
dtype: string
- name: token_type_ids
dtype: string
- name: page_id
dtype: int64
splits:
- name: train
num_bytes: 969109
num_examples: 30
- name: validation
num_bytes: 1086479
num_examples: 29
download_size: 1838074
dataset_size: 2055588
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*