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: 27715208700
num_examples: 2253269
download_size: 7902918885
dataset_size: 27715208700