metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: label
dtype: string
- name: original_file_name
dtype: string
- name: audio_embeddings
sequence: float32
splits:
- name: train
num_bytes: 3023882524.25
num_examples: 17206
- name: validation
num_bytes: 373249770.75
num_examples: 2194
- name: test
num_bytes: 369753644
num_examples: 2144
download_size: 2808036364
dataset_size: 3766885939
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*