metadata
dataset_info:
features:
- name: data_combined
sequence: int64
- name: retrieve_start_idx_list
sequence: int64
- name: retrieve_score_list
sequence: float64
splits:
- name: train
num_bytes: 3697624044
num_examples: 53779
- name: val
num_bytes: 206312448
num_examples: 3000
download_size: 596762594
dataset_size: 3903936492
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*