metadata
dataset_info:
features:
- name: id
dtype: int64
- name: audio
dtype: string
- name: conversation
list:
- name: from
dtype: string
- name: value
dtype: string
- name: language
dtype: string
- name: pretraining_split
dtype: bool
splits:
- name: train
num_bytes: 2858691549
num_examples: 9337075
- name: test
num_bytes: 323721.5
num_examples: 1000
download_size: 1315993522
dataset_size: 2859015270.5
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*