metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 5250716600
num_examples: 787450
- name: validation
num_bytes: 187457484
num_examples: 28113
download_size: 1353665677
dataset_size: 5438174084
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*