metadata
dataset_info:
features:
- name: speaker
dtype: string
- name: start
dtype: int64
- name: end
dtype: int64
- name: text
dtype: string
- name: dialogue_id
dtype: string
- name: turn_id
dtype: int64
- name: original_split
dtype: string
splits:
- name: train
num_bytes: 36015874
num_examples: 272994
download_size: 13936433
dataset_size: 36015874
configs:
- config_name: default
data_files:
- split: train
path: data/train-*