metadata
dataset_info:
features:
- name: npy_bytes
dtype: binary
- name: png_image
dtype: image
- name: relative_path
dtype: string
- name: relative_dir
dtype: string
- name: filename
dtype: string
- name: file_type
dtype: string
- name: file_size
dtype: int64
splits:
- name: train
num_bytes: 39885201198
num_examples: 806412
download_size: 45247046566
dataset_size: 39885201198
configs:
- config_name: default
data_files:
- split: train
path: data/train-*