metadata
dataset_info:
features:
- name: doc_id
dtype: string
- name: tokens
list: string
- name: bboxes
list:
list: int64
- name: query
dtype: string
- name: query_type
dtype: string
- name: labels
list: string
- name: table_labels
list: int64
- name: row_labels
list: int64
- name: is_negative
dtype: bool
splits:
- name: train
num_bytes: 2031394758
num_examples: 180009
download_size: 24739260
dataset_size: 2031394758
configs:
- config_name: default
data_files:
- split: train
path: data/train-*