metadata
dataset_info:
features:
- name: speaker
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: embeddings
dtype:
array2_d:
shape:
- 1
- 256
dtype: float32
splits:
- name: train
num_bytes: 261482
num_examples: 2
download_size: 212141
dataset_size: 261482
configs:
- config_name: default
data_files:
- split: train
path: data/train-*