metadata
dataset_info:
features:
- name: id
dtype: string
- name: input_text
sequence: string
- name: start_positions
dtype: int64
- name: end_positions
dtype: int64
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 500267402
num_examples: 130318
- name: validation
num_bytes: 46561262
num_examples: 11873
download_size: 64536033
dataset_size: 546828664
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*