metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: input
dtype: string
- name: output
sequence: string
splits:
- name: train
num_bytes: 1646765
num_examples: 2768
- name: test
num_bytes: 1648925
num_examples: 2767
- name: dev
num_bytes: 68696
num_examples: 116
download_size: 354595
dataset_size: 3364386
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: dev
path: data/dev-*