metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: input
dtype: 'null'
- name: output
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 25003580
num_examples: 11889
download_size: 8187105
dataset_size: 25003580
configs:
- config_name: default
data_files:
- split: train
path: data/train-*