metadata
dataset_info:
features:
- name: inputs
dtype: string
- name: targets
dtype: string
- name: _template_idx
dtype: int64
- name: _task_source
dtype: string
- name: _task_name
dtype: string
- name: _template_type
dtype: string
splits:
- name: train
num_bytes: 66678959
num_examples: 85415
download_size: 32544242
dataset_size: 66678959
configs:
- config_name: default
data_files:
- split: train
path: data/train-*