metadata
dataset_info:
features:
- name: caption
dtype: string
- name: jpg_0
dtype: binary
- name: jpg_1
dtype: binary
- name: label_0
dtype: int64
- name: label_1
dtype: int64
- name: has_label
dtype: bool
- name: are_different
dtype: bool
splits:
- name: train
num_bytes: 231731781
num_examples: 2000
download_size: 230141961
dataset_size: 231731781
configs:
- config_name: default
data_files:
- split: train
path: data/train-*