metadata
dataset_info:
features:
- name: pixel_values
sequence:
sequence:
sequence: float32
- name: label
dtype: int64
splits:
- name: train
num_bytes: 14337352920
num_examples: 23705
download_size: 3114082212
dataset_size: 14337352920
configs:
- config_name: default
data_files:
- split: train
path: data/train-*