metadata
dataset_info:
features:
- name: conversation
list:
- name: content
dtype: string
- name: language
dtype: string
- name: role
dtype: string
- name: asr
dtype: bool
- name: prompt_lang
dtype: string
- name: answer_lang
dtype: string
- name: index
dtype: int64
splits:
- name: train
num_bytes: 78089091
num_examples: 39308
download_size: 37098392
dataset_size: 78089091
configs:
- config_name: default
data_files:
- split: train
path: data/train-*