metadata
dataset_info:
features:
- name: question
dtype: string
- name: answer
dtype: string
- name: tokenized_question
dtype: string
- name: tokenized_answer
dtype: string
- name: context
dtype: string
- name: tokenized_context
dtype: string
- name: negatives
sequence: string
- name: tokenized_negatives
sequence: string
- name: id
dtype: string
- name: candidates
sequence: string
- name: candidates_tokenized
sequence: string
splits:
- name: train
num_bytes: 91555125
num_examples: 5829
download_size: 30725565
dataset_size: 91555125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*