metadata
dataset_info:
features:
- name: metadata
struct:
- name: doc_id
dtype: string
- name: split
dtype: int64
- name: title
dtype: string
- name: token_count
dtype: int64
- name: url
dtype: string
- name: passage
dtype: string
- name: id
dtype: int64
splits:
- name: train
num_bytes: 1377054502
num_examples: 1639166
download_size: 605204760
dataset_size: 1377054502
configs:
- config_name: default
data_files:
- split: train
path: data/train-*