metadata
dataset_info:
features:
- name: context
sequence: string
- name: question
sequence: string
- name: answer_text
dtype: string
- name: answer_start_idx
dtype: int64
- name: label
dtype: int64
- name: answer_word_start_idx
dtype: int64
- name: answer_word_end_idx
dtype: int64
- name: input_ids
sequence: int32
- name: words_lengths
sequence: int64
- name: start_idx
dtype: int64
- name: end_idx
dtype: int64
- name: valid
dtype: bool
splits:
- name: train
num_bytes: 98164339.55359063
num_examples: 15354
download_size: 19264083
dataset_size: 98164339.55359063
configs:
- config_name: default
data_files:
- split: train
path: data/train-*