metadata
dataset_info:
features:
- name: image
dtype: image
- name: bboxes
sequence:
sequence: float64
- name: category_id
sequence: int64
- name: input_text
dtype: string
- name: output_text
dtype: string
splits:
- name: train
num_bytes: 27278694249.125
num_examples: 69375
- name: validation
num_bytes: 2408827639.875
num_examples: 6489
- name: test
num_bytes: 1926932272.125
num_examples: 4999
download_size: 31483689922
dataset_size: 31614454161.125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*