metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 35176479216
num_examples: 2640084
- name: validation
num_bytes: 1845267408
num_examples: 138492
download_size: 11034181883
dataset_size: 37021746624
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*