metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: word_ids
sequence: int64
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 5380537328
num_examples: 991987
download_size: 1379723543
dataset_size: 5380537328