metadata
dataset_info:
features:
- name: img1
dtype: string
- name: img2
dtype: string
- name: label
dtype: int64
splits:
- name: train
num_bytes: 366236
num_examples: 6000
download_size: 159715
dataset_size: 366236
configs:
- config_name: default
data_files:
- split: train
path: data/train-*