metadata
dataset_info:
- config_name: corpus
features:
- name: docid
dtype: string
- name: title
dtype: string
- name: text
dtype: string
splits:
- name: train
num_bytes: 927468420
num_examples: 325427
download_size: 560028392
dataset_size: 927468420
- config_name: default
features:
- name: query_id
dtype: string
- name: query
dtype: string
- name: positive_passages
list:
- name: docid
dtype: string
- name: text
dtype: string
- name: title
dtype: string
- name: hard_negative_passages
sequence: 'null'
- name: negative_passages
list:
- name: docid
dtype: string
- name: text
dtype: string
- name: title
dtype: string
- name: metadata
struct:
- name: domain
dtype: string
- name: template
dtype: string
splits:
- name: train
num_bytes: 429323533
num_examples: 1297
- name: train_aug
num_bytes: 1769945045
num_examples: 6297
- name: val
num_bytes: 105711070
num_examples: 323
- name: test
num_bytes: 86241005
num_examples: 1727
download_size: 1439017940
dataset_size: 2391220653
configs:
- config_name: corpus
data_files:
- split: train
path: corpus/train-*
- config_name: default
data_files:
- split: train
path: data/train-*
- split: train_aug
path: data/train_aug-*
- split: val
path: data/val-*
- split: test
path: data/test-*