metadata
dataset_info:
features:
- name: query
dtype: string
- name: documents
sequence: string
- name: scores
sequence: float64
splits:
- name: train
num_bytes: 2054802402
num_examples: 138384
download_size: 1202834615
dataset_size: 2054802402
configs:
- config_name: default
data_files:
- split: train
path: data/train-*