metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: type
dtype: string
- name: instruction
dtype: string
- name: input
sequence: string
- name: output
struct:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 56299970
num_examples: 31516
download_size: 29783446
dataset_size: 56299970
configs:
- config_name: default
data_files:
- split: train
path: data/train-*