metadata
dataset_info:
features:
- name: image
dtype: image
- name: contest_number
dtype: int64
- name: captions
sequence: string
- name: mean
sequence: float64
splits:
- name: train
num_bytes: 359384521
num_examples: 365
download_size: 294881876
dataset_size: 359384521
configs:
- config_name: default
data_files:
- split: train
path: data/train-*