metadata
dataset_info:
features:
- name: input_ids_1
sequence: int64
- name: att_1
sequence: int64
- name: query
dtype: string
- name: doc_1
dtype: string
- name: doc_2
dtype: string
- name: input_ids_2
sequence: int64
- name: att_2
sequence: int64
- name: label
dtype: int64
- name: pair_id
dtype: int64
- name: pred
dtype: int64
splits:
- name: train
num_bytes: 49610561
num_examples: 2766
download_size: 2469477
dataset_size: 49610561
configs:
- config_name: default
data_files:
- split: train
path: data/train-*