metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: validation
num_bytes: 1366446672
num_examples: 33354
- name: train
num_bytes: 135315747216
num_examples: 3302962
download_size: 53030280443
dataset_size: 136682193888
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*
- split: train
path: data/train-*