metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption
dtype: string
- name: label
dtype: int64
splits:
- name: train
num_bytes: 65841507
num_examples: 512
download_size: 65837860
dataset_size: 65841507
configs:
- config_name: default
data_files:
- split: train
path: data/train-*