metadata
dataset_info:
features:
- name: id
dtype: string
- name: conv_A
list:
- name: role
dtype: string
- name: content
dtype: string
- name: conv_B
list:
- name: role
dtype: string
- name: content
dtype: string
- name: conv_A_rating
dtype: float32
- name: conv_B_rating
dtype: float32
- name: num_turns
dtype: int32
- name: source
dtype: string
splits:
- name: train
num_bytes: 97672090
num_examples: 25386
- name: val
num_bytes: 3796982
num_examples: 1000
download_size: 54581615
dataset_size: 101469072
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*