metadata
dataset_info:
features:
- name: image
dtype: image
- name: image_id
dtype: int64
- name: new_image_id
dtype: int64
- name: answers
sequence: string
splits:
- name: train
num_bytes: 8209032325
num_examples: 316219
download_size: 8106455712
dataset_size: 8209032325
configs:
- config_name: default
data_files:
- split: train
path: data/train-*