metadata
dataset_info:
features:
- name: task_id
dtype: string
- name: image
list:
list: int64
splits:
- name: train
num_bytes: 13218000
num_examples: 400
download_size: 82074
dataset_size: 13218000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*