metadata
dataset_info:
features:
- name: instruction
struct:
- name: role
dtype: string
- name: content
dtype: string
- name: input
list:
- name: role
dtype: string
- name: content
dtype: string
- name: output
dtype: string
- name: input_ids
sequence: int32
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 2642247.214285714
num_examples: 13
- name: test
num_bytes: 100362
num_examples: 1
download_size: 28214
dataset_size: 2742609.214285714
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*