metadata
dataset_info:
features:
- name: speaker_name
dtype: string
- name: start_time
dtype: float64
- name: end_time
dtype: float64
- name: caption_text
dtype: string
- name: episode_id
dtype: int64
- name: speaker_id
dtype: int64
- name: clip_id
dtype: int64
- name: audio
dtype:
audio:
sampling_rate: 16000
splits:
- name: train
num_bytes: 349869882.712
num_examples: 6288
- name: val
num_bytes: 131023230.92
num_examples: 2080
download_size: 472289209
dataset_size: 480893113.632
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*