metadata
dataset_info:
config_name: truncate-512
features:
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: special_tokens_mask
sequence: int8
splits:
- name: train
num_bytes: 23600541600
num_examples: 6555706
- name: validation
num_bytes: 317304000
num_examples: 88140
download_size: 310440269
dataset_size: 23917845600
configs:
- config_name: truncate-512
data_files:
- split: train
path: truncate-512/train-*
- split: validation
path: truncate-512/validation-*