metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 9023944428
num_examples: 1353321
- name: valid
num_bytes: 263452680
num_examples: 39510
download_size: 154167252
dataset_size: 9287397108
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*