metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: input
dtype: string
- name: response
dtype: string
splits:
- name: train
num_bytes: 10931600.8
num_examples: 8000
- name: valid
num_bytes: 1366450.1
num_examples: 1000
- name: test
num_bytes: 1366450.1
num_examples: 1000
download_size: 2008085
dataset_size: 13664501
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*
- split: test
path: data/test-*