metadata
dataset_info:
features:
- name: text
dtype: string
- name: domain
list: string
- name: hal_id
dtype: string
- name: url
dtype: string
- name: year
dtype: string
- name: token_count
dtype: int64
splits:
- name: train
num_bytes: 3488116389
num_examples: 23299
download_size: 1741406846
dataset_size: 3488116389
configs:
- config_name: default
data_files:
- split: train
path: data/train-*