metadata
dataset_info:
features:
- name: gt_blocks
dtype: string
- name: img
dtype: image
- name: page_bbox
sequence: float32
- name: classification
dtype: string
- name: language
dtype: string
- name: pdf
dtype: binary
- name: uuid
dtype: int64
splits:
- name: train
num_bytes: 775967309.75
num_examples: 2138
download_size: 760821924
dataset_size: 775967309.75
configs:
- config_name: default
data_files:
- split: train
path: data/train-*