metadata
dataset_info:
features:
- name: input_ids
list: int32
- name: attention_mask
list: int8
- name: labels
list: int64
splits:
- name: train
num_bytes: 7170915885
num_examples: 14390566
- name: test
num_bytes: 2860241
num_examples: 6043
- name: valid
num_bytes: 3264678
num_examples: 6492
download_size: 10314145660
dataset_size: 7177040804
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*