metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 17609664600
num_examples: 1321650
- name: validation
num_bytes: 914479416
num_examples: 68634
download_size: 5516198279
dataset_size: 18524144016
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*