metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: image
dtype: image
- name: description
dtype: string
- name: description_tokens
dtype: int64
splits:
- name: train
num_bytes: 77603794659
num_examples: 58501
download_size: 77509690222
dataset_size: 77603794659