metadata
dataset_info:
features:
- name: text
dtype: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 5998238715
num_examples: 2119719
- name: validation
num_bytes: 61703030
num_examples: 21990
download_size: 1771825341
dataset_size: 6059941745
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*