metadata
dataset_info:
features:
- name: query
struct:
- name: title
dtype: string
- name: abstract
dtype: string
- name: corpus_id
dtype: uint64
- name: pos
struct:
- name: title
dtype: string
- name: abstract
dtype: string
- name: corpus_id
dtype: uint64
- name: neg
struct:
- name: title
dtype: string
- name: abstract
dtype: string
- name: corpus_id
dtype: uint64
splits:
- name: train
num_bytes: 2069439948
num_examples: 475656
download_size: 1222811951
dataset_size: 2069439948
configs:
- config_name: default
data_files:
- split: train
path: data/train-*