metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 1767763064
num_examples: 2094506
- name: test
num_bytes: 162285164
num_examples: 192281
download_size: 239611984
dataset_size: 1930048228
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*