metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: speaker_id
dtype: string
- name: wavlm_embedding
list: float64
splits:
- name: train
num_bytes: 2861801092
num_examples: 4500
download_size: 2646894701
dataset_size: 2861801092
configs:
- config_name: default
data_files:
- split: train
path: data/train-*