metadata
dataset_info:
features:
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 928847774.5358719
num_examples: 7635694
- name: test
num_bytes: 1216454.9476915547
num_examples: 10000
download_size: 523969256
dataset_size: 930064229.4835634
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
256トークンで切った。