metadata
dataset_info:
features:
- name: input
dtype: string
- name: instruction
dtype: string
- name: response
dtype: string
splits:
- name: train
num_bytes: 103993344739
num_examples: 78131251
download_size: 27852811712
dataset_size: 103993344739
configs:
- config_name: default
data_files:
- split: train
path: data/train-*