metadata
dataset_info:
features:
- name: id
dtype: string
- name: prompt
dtype: string
- name: answers
struct:
- name: text
sequence: string
- name: answer_start
sequence: int64
splits:
- name: train
num_bytes: 4006163
num_examples: 3000
- name: validation
num_bytes: 4006163
num_examples: 3000
download_size: 1496526
dataset_size: 8012326
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*