metadata
dataset_info:
features:
- name: relation
dtype: string
- name: query
dtype: string
- name: subject
dtype: string
- name: object
dtype: string
- name: template
dtype: string
- name: template_index
dtype: int64
- name: candidates
sequence: string
splits:
- name: train
num_bytes: 293951032
num_examples: 157656
download_size: 8219151
dataset_size: 293951032
configs:
- config_name: default
data_files:
- split: train
path: data/train-*