metadata
dataset_info:
features:
- name: context
dtype: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 1766922390
num_examples: 20000
download_size: 592965039
dataset_size: 1766922390