metadata
license: apache-2.0
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 93845075764
num_examples: 7043311
- name: validation
num_bytes: 9126940
num_examples: 685
download_size: 27612151408
dataset_size: 93854202704
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*