metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 84157459
num_examples: 20970
download_size: 47243194
dataset_size: 84157459
configs:
- config_name: default
data_files:
- split: train
path: data/train-*