metadata
dataset_info:
features:
- name: image
dtype: image
- name: bboxes
sequence:
sequence: int64
- name: text
sequence: string
- name: language
dtype: string
splits:
- name: train
num_bytes: 478482327.362134
num_examples: 4635
download_size: 436363110
dataset_size: 478482327.362134
configs:
- config_name: default
data_files:
- split: train
path: data/train-*