metadata
dataset_info:
features:
- name: image_url
dtype: string
- name: caption
dtype: string
- name: transcripts
list: string
splits:
- name: train
num_bytes: 19950157
num_examples: 7170
download_size: 16695542
dataset_size: 19950157
configs:
- config_name: default
data_files:
- split: train
path: data/train-*