metadata
dataset_info:
features:
- name: file
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 22050
- name: genre
dtype:
class_label:
names:
'0': blues
'1': classical
'2': country
'3': disco
'4': hiphop
'5': jazz
'6': metal
'7': pop
'8': reggae
'9': rock
splits:
- name: train
num_bytes: 1190490124.4264264
num_examples: 899
- name: test
num_bytes: 132458065.57357357
num_examples: 100
download_size: 1321418545
dataset_size: 1322948190
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
from marsyas/gtzan, split it into train and test sets randomly