metadata
dataset_info:
features:
- name: image_id
dtype: int32
- name: alt_text
dtype: string
- name: image
dtype: image
- name: source
dtype: string
splits:
- name: train
num_bytes: 4184259
num_examples: 100
download_size: 4175096
dataset_size: 4184259
configs:
- config_name: default
data_files:
- split: train
path: data/train-*