Skip to content

sine_wave

Sine Wave Layer API

This module provides classes to build sine wave layers.

Classes:

Classes

AddSineWave 

AddSineWave(sample_rate: float = 1, frequency: float = 100, amplitude: float = 0.1, data_format: str | None = None, **kwargs)

Adds a sine wave to the input.

Parameters:

  • sample_rate

    (float, default: 1 ) –

    Sample rate of the input.

  • frequency

    (float, default: 100 ) –

    Frequency of the sine wave in Hz.

  • amplitude

    (float, default: 0.1 ) –

    Amplitude of the sine wave.

  • data_format

    (str, default: None ) –

    Data format of the input.

Source code in neuralspot_edge/layers/preprocessing/sine_wave.py 
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
def __init__(
    self,
    sample_rate: float = 1,
    frequency: float = 100,
    amplitude: float = 0.1,
    data_format: str | None = None,
    **kwargs,
):
    """Adds a sine wave to the input.

    Args:
        sample_rate (float): Sample rate of the input.
        frequency (float): Frequency of the sine wave in Hz.
        amplitude (float): Amplitude of the sine wave.
        data_format (str): Data format of the input.
    """
    super().__init__(**kwargs)
    if sample_rate < 0:
        raise ValueError("sample_rate must be greater than 0")
    if 0 > frequency or frequency > sample_rate / 2:
        raise ValueError(" 0 <= frequency < sample_rate / 2")
    self.sample_rate = sample_rate
    self.frequency = frequency
    self.amplitude = amplitude
    self.data_format = data_format or keras.backend.image_data_format()

    if self.data_format == "channels_first":
        self.duration_axis = -1
        self.ch_axis = -2
    else:
        self.duration_axis = -2
        self.ch_axis = -1

Functions

call 
call(inputs, training=True)

Augment single sample with sine wave.

Source code in neuralspot_edge/layers/preprocessing/sine_wave.py 
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
def call(self, inputs, training=True):
    """Augment single sample with sine wave."""
    self.training = training
    is_batched = inputs.shape.rank == 3
    if not is_batched:
        inputs = keras.ops.expand_dims(inputs, axis=0)
    # END
    if self.training:
        input_shape = keras.ops.shape(inputs)
        batch_size = input_shape[0]
        duration_size = input_shape[self.duration_axis]
        ch_size = input_shape[self.ch_axis]

        # Create sine wave at the specified frequency using sample_rate as the time step
        ts = keras.ops.cast(keras.ops.arange(duration_size), self.compute_dtype) / self.sample_rate

        sine_wave = keras.ops.sin(2 * np.pi * self.frequency * ts)
        if self.data_format == "channels_first":
            sine_wave = keras.ops.reshape(sine_wave, (1, 1, duration_size))
            sine_wave = keras.ops.tile(sine_wave, (batch_size, ch_size, 1))
        else:
            sine_wave = keras.ops.reshape(sine_wave, (1, duration_size, 1))
            sine_wave = keras.ops.tile(sine_wave, (batch_size, 1, ch_size))
        outputs = keras.ops.cast(inputs + self.amplitude * sine_wave, self.compute_dtype)
    # END IF

    if not is_batched:
        outputs = keras.ops.squeeze(outputs, axis=0)
    # END IF
    return outputs