reservoirpy.datasets.japanese_vowels#

reservoirpy.datasets.japanese_vowels(one_hot_encode=True, repeat_targets=False, data_folder=None, reload=False)[source]#

Load the Japanese vowels [16] dataset.

This is a classic audio discrimination task. Nine male Japanese speakers pronounced the ` ae` vowel. The task consists in inferring the speaker identity from the audio recording.

Audio recordings are series of 12 LPC cepstrum coefficient. Series contains between 7 and 29 timesteps. Each series (or “block”) is one utterance of ` ae` vowel from one speaker.

Classes

9

Samples per class (training)

30 series of 7-29 timesteps

Samples per class (testing)

29-50 series of 7-29 timesteps

Samples total

640

Dimensionality

12

Features

real

Data is downloaded from: https://doi.org/10.24432/C5NS47

Parameters:

  • one_hot_encode (bool, default to True) – If True, returns class label as a one-hot encoded vector.

  • repeat_targets (bool, default to False) – If True, repeat the target label or vector along the time axis of the corresponding sample.

  • data_folder (str or Path-like object, optional) – Local destination of the downloaded data.

  • reload (bool, default to False) – If True, re-download data from remote repository. Else, if a cached version of the dataset exists, use the cached dataset.

Returns:

Lists of arrays of shape (timesteps, features) or (timesteps, target) or (target,).

Return type:

X_train, Y_train, X_test, Y_test

References