metadata
dataset_info:
features:
- name: url
dtype: string
- name: caption
sequence: string
splits:
- name: train
num_bytes: 40997825
num_examples: 118287
download_size: 18819985
dataset_size: 40997825
configs:
- config_name: default
data_files:
- split: train
path: data/train-*