metadata
dataset_info:
features:
- name: inputs
dtype: string
- name: targets
dtype: string
- name: task
dtype: string
- name: target_token_count
dtype: int64
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 10279200.81498829
num_examples: 4302
download_size: 5158288
dataset_size: 10279200.81498829
configs:
- config_name: default
data_files:
- split: train
path: data/train-*