metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: validation
path: data/validation-*
dataset_info:
features:
- name: pixel_values
sequence:
sequence:
sequence: float32
- name: labels
sequence:
sequence: int8
splits:
- name: train
num_bytes: 5903889992
num_examples: 12551
- name: test
num_bytes: 959599680
num_examples: 2040
- name: validation
num_bytes: 1537241056
num_examples: 3268
download_size: 4679673366
dataset_size: 8400730728