metadata
dataset_info:
features:
- name: context
dtype: string
- name: question
dtype: string
- name: answer
struct:
- name: answer_end
dtype: int64
- name: answer_start
dtype: int64
- name: text
dtype: string
- name: input_ids
sequence: int32
- name: start_positions
dtype: int64
- name: end_positions
dtype: int64
splits:
- name: train
num_bytes: 4286291
num_examples: 2976
- name: test
num_bytes: 703898
num_examples: 478
download_size: 1197807
dataset_size: 4990189
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*