metadata
dataset_info:
features:
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: sft_template_type_idx
dtype: string
splits:
- name: train
num_bytes: 326805821
num_examples: 54450
download_size: 122447738
dataset_size: 326805821
configs:
- config_name: default
data_files:
- split: train
path: data/train-*