metadata
dataset_info:
features:
- name: token
sequence:
sequence: int16
- name: label
dtype: int64
splits:
- name: val
num_bytes: 109400000
num_examples: 50000
- name: train
num_bytes: 2803193396
num_examples: 1281167
download_size: 1543647217
dataset_size: 2912593396
configs:
- config_name: default
data_files:
- split: val
path: data/val-*
- split: train
path: data/train-*