metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: response
dtype: string
splits:
- name: train
num_bytes: 2479736208
num_examples: 645000
download_size: 914598500
dataset_size: 2479736208
configs:
- config_name: default
data_files:
- split: train
path: data/train-*