metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: next_sentence_label
dtype: int64
splits:
- name: train
num_bytes: 70985500
num_examples: 2000
download_size: 18506503
dataset_size: 70985500