metadata
dataset_info:
features:
- name: image_id
dtype: string
- name: image
dtype: image
- name: width
dtype: int64
- name: height
dtype: int64
- name: tokens
sequence: string
- name: bboxes
sequence:
sequence: float64
- name: labels
sequence: string
splits:
- name: train
num_bytes: 316972972
num_examples: 1440
- name: test
num_bytes: 22266294
num_examples: 100
download_size: 680849388
dataset_size: 339239266
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*