random_background_noises

Name	Description
`RandomBackgroundNoises1D`	Random background noises 1D

Name

Description

RandomBackgroundNoises1D

Random background noises 1D

RandomBackgroundNoises1D(noises, amplitude: float | tuple[float, float] = 0.1, num_noises: int = 1, **kwargs={})

Apply random background noises to the input.

Example:

    sample_rate = 100
    duration = 2*sample_rate
    freqs = [2, 5, 15, 20, 25]
    noises = np.vstack([
        np.sin(2*np.pi*f*np.arange(duration)/sample_rate)
        for f in freqs
    ]).T
    lyr = RandomBackgroundNoises(noises=noises, amplitude=0.2, num_noises=2)
    y = lyr(x, training=True)

Base class: BaseAugmentation1D.

Inherited from BaseAugmentation: augment_masks(), augment_samples(), augment_targets(), batch_augment(), call().

Parameters

Parameters of RandomBackgroundNoises1D
Name	Type	Default	Description
`noises`	`np.ndarray`	Required	Background noises to apply.
`amplitude`	`float \| tuple[float, float]`	0.1	Amplitude of the noise. If tuple, amplitude is randomly picked between the values.

amplitude: tuple[float, float] = parse_factor(amplitude, min_value=0, max_value=None, param_name='amplitude')

num_noises: int = num_noises

noises = self.add_weight(name='noises', shape=noises.shape, initializer=keras.initializers.Constant(noises), trainable=False)

get_random_transformations(input_shape: tuple[int, int, int]) -> dict

Generate noise tensor

Parameters

Parameters of get_random_transformations
Name	Type	Default	Description
`input_shape`	`tuple[int, ...]`	Required	Input shape.

Returns

Returns of get_random_transformations
Value	Type	Description
`dict`	`dict`	Dictionary containing the noise tensor.

augment_sample(inputs) -> keras.KerasTensor

Augment single sample with random background noises.

get_config()

Serializes the configuration of the layer.

random_background_noises

Random Background Noises Layer API