metadata
dataset_info:
features:
- name: image
dtype: string
- name: relationships_text
dtype: string
- name: objects
sequence: string
splits:
- name: train
num_bytes: 7066
num_examples: 3
download_size: 18856
dataset_size: 7066
configs:
- config_name: default
data_files:
- split: train
path: data/train-*