metadata
dataset_info:
features:
- name: input
dtype: string
- name: output
dtype: int64
- name: a
dtype: int64
- name: b
dtype: int64
splits:
- name: train
num_bytes: 587690240
num_examples: 10000000
- name: test
num_bytes: 587464
num_examples: 10000
download_size: 206168581
dataset_size: 588277704
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*