tf.keras.layers.GaussianNoise(
stddev, **kwargs
)
description
Apply Gaussian noise.
This is useful for reducing overfitting (think of it as a form of random data augmentation). Gaussian noise (GS) is a natural choice as the destruction process of real-valued input.
Since it is a regularization layer, it is only activated during training
parameter
stddev
floating point number, the standard deviation of the noise distribution
Callable parameters
inputs
input tensor
training
Boolean value, indicating whether the layer should be run in training mode (add noise) or in inference mode (do nothing)
Input shape
When using this layer as the first layer of the model, use the keyword parameter input_shape
Output shape
Output with the same shape as the input