metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: special_tokens_mask
sequence: int8
splits:
- name: train
num_bytes: 7023324732
num_examples: 2277358
- name: validation
num_bytes: 372257304
num_examples: 120706
download_size: 2053391840
dataset_size: 7395582036
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*