metadata
dataset_info:
features:
- name: image
dtype: image
- name: image_id
dtype: string
- name: input
dtype: string
- name: caption
dtype: string
splits:
- name: train
num_bytes: 18442237
num_examples: 198
download_size: 7499940
dataset_size: 18442237
configs:
- config_name: default
data_files:
- split: train
path: data/train-*