metadata
dataset_info:
features:
- name: input_ids
sequence: int64
- name: attention_mask
sequence: int64
- name: answer
dtype: string
- name: start_positions
dtype: int64
- name: end_positions
dtype: int64
splits:
- name: train
num_bytes: 414807
num_examples: 50
download_size: 50631
dataset_size: 414807
configs:
- config_name: default
data_files:
- split: train
path: data/train-*