metadata
dataset_info:
features:
- name: image
dtype: image
- name: objects
struct:
- name: Ids
sequence: string
- name: captions
sequence: string
- name: quantities
sequence: int64
splits:
- name: train
num_bytes: 1221170619.216
num_examples: 20244
download_size: 1142033992
dataset_size: 1221170619.216
configs:
- config_name: default
data_files:
- split: train
path: data/train-*