metadata
dataset_info:
features:
- name: source
dtype: string
- name: caption_ko
sequence: string
splits:
- name: train
num_bytes: 45203310
num_examples: 122256
download_size: 20493596
dataset_size: 45203310
configs:
- config_name: default
data_files:
- split: train
path: data/train-*