metadata
dataset_info:
features:
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
splits:
- name: train
num_bytes: 11201825482
num_examples: 500000
- name: validation
num_bytes: 45438138
num_examples: 2000
download_size: 11246655245
dataset_size: 11247263620
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*