metadata
dataset_info:
features:
- name: task_id
dtype: string
- name: train
list:
- name: input
list:
list: int64
- name: output
list:
list: int64
- name: test_input
list:
list: int64
- name: test_output
list:
list: int64
splits:
- name: train
num_bytes: 3631364
num_examples: 386
- name: evaluation
num_bytes: 5994736
num_examples: 381
download_size: 340716
dataset_size: 9626100
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: evaluation
path: data/evaluation-*
ARC AGI Dataset
