metadata
dataset_info:
features:
- name: image
dtype: image
- name: bboxes
sequence:
sequence: float64
- name: labels
sequence: int64
- name: words
sequence: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 32034973965.125
num_examples: 80863
download_size: 0
dataset_size: 32034973965.125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*