metadata
dataset_info:
features:
- name: ref
dtype: string
- name: title
dtype: string
- name: mtf
dtype: string
- name: abc
dtype: string
splits:
- name: train
num_bytes: 1673713
num_examples: 663
download_size: 811890
dataset_size: 1673713
configs:
- config_name: default
data_files:
- split: train
path: data/train-*