metadata
dataset_info:
features:
- name: text_chunk
sequence: int64
- name: predicted_token
dtype: int64
splits:
- name: train
num_bytes: 2214065356
num_examples: 12872473
download_size: 253197513
dataset_size: 2214065356
configs:
- config_name: default
data_files:
- split: train
path: data/train-*