metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption
dtype: string
- name: image_file
dtype: string
splits:
- name: train
num_bytes: 16819843770.21
num_examples: 125494
download_size: 6179000593
dataset_size: 16819843770.21
configs:
- config_name: default
data_files:
- split: train
path: data/train-*