metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: query
dtype: string
- name: pos
sequence: string
- name: neg
sequence: string
- name: task
dtype: string
- name: instruction
struct:
- name: query
dtype: string
- name: pos
dtype: string
- name: neg
dtype: string
splits:
- name: train
num_bytes: 848430524
num_examples: 200000
download_size: 523334138
dataset_size: 848430524