metadata
dataset_info:
features:
- name: doc_id
dtype: string
- name: split
dtype: string
- name: doc_type
dtype: string
- name: raw_pdf_path
dtype: string
- name: label_path
dtype: string
- name: page_image_path
dtype: string
- name: page_num
dtype: int64
- name: page_ocr_text
dtype: string
- name: page_pdf_text
dtype: string
- name: visual_elements
dtype: string
- name: table_elements
dtype: string
- name: ocr_tokens
dtype: int64
- name: pdf_tokens
dtype: int64
- name: table_len
dtype: int64
- name: image_len
dtype: int64
splits:
- name: train
num_bytes: 159866003
num_examples: 38667
- name: valid
num_bytes: 20656820
num_examples: 4904
download_size: 69048503
dataset_size: 180522823
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*