metadata
dataset_info:
features:
- name: tokens
sequence: int64
- name: targets
sequence: int64
splits:
- name: train
num_bytes: 49288800
num_examples: 3300
download_size: 3044454
dataset_size: 49288800
configs:
- config_name: default
data_files:
- split: train
path: data/train-*