metadata
dataset_info:
features:
- name: task_name
dtype: string
- name: id
dtype: string
- name: definition
dtype: string
- name: inputs
dtype: string
- name: targets
dtype: string
splits:
- name: train
num_bytes: 83627494
num_examples: 75636
download_size: 22971372
dataset_size: 83627494
configs:
- config_name: default
data_files:
- split: train
path: data/train-*