metadata
dataset_info:
features:
- name: image
dtype: image
- name: bboxes
sequence:
sequence: int32
- name: labels
sequence: int32
- name: words
sequence: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 1076570989.75
num_examples: 2890
download_size: 1065785831
dataset_size: 1076570989.75
configs:
- config_name: default
data_files:
- split: train
path: data/train-*