metadata
dataset_info:
features:
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: source
dtype: string
- name: conversation_id
dtype: int64
splits:
- name: train
num_bytes: 3586592651
num_examples: 2172778
download_size: 1723637483
dataset_size: 3586592651
configs:
- config_name: default
data_files:
- split: train
path: data/train-*