metadata
dataset_info:
features:
- name: conversation_id
dtype: string
- name: topic
dtype: string
- name: turn_id
dtype: int64
- name: has_content
dtype: bool
- name: input_history
list:
- name: audio_file
dtype: string
- name: content
dtype: string
- name: role
dtype: string
- name: audio_file
dtype: string
- name: user
dtype: string
- name: assistant
dtype: string
splits:
- name: train
num_bytes: 1112009
num_examples: 340
- name: test
num_bytes: 1129399
num_examples: 250
- name: validation
num_bytes: 59468
num_examples: 50
download_size: 192053
dataset_size: 2300876
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: validation
path: data/validation-*