metadata
language:
- en
dataset_info:
features:
- name: id
dtype: string
- name: topdown
dtype: image
- name: choices
dtype: image
- name: up
dtype: image
- name: right
dtype: image
- name: down
dtype: image
- name: left
dtype: image
- name: answer
dtype: string
splits:
- name: train
num_bytes: 183285910
num_examples: 250
download_size: 183297503
dataset_size: 183285910
configs:
- config_name: default
data_files:
- split: train
path: data/train-*