metadata
dataset_info:
features:
- name: query
dtype: string
- name: document
dtype: string
- name: hard_negative
dtype: string
- name: query_document_similarity_score
dtype: float64
- name: query_negative_similarity_score
dtype: float64
splits:
- name: train
num_bytes: 640233555
num_examples: 744862
download_size: 280508154
dataset_size: 640233555
configs:
- config_name: default
data_files:
- split: train
path: data/train-*