Contents

init(type:beta:gamma:momentum:epsilon:activation:)

Returns a new fused normalization parameters structure.

Declaration

init(type: BNNS.NormalizationType, beta: BNNSNDArrayDescriptor? = nil, gamma: BNNSNDArrayDescriptor? = nil, momentum: Float = 0, epsilon: Float, activation: BNNS.ActivationFunction)

Parameters

  • type:

    An enumeration that specifies the normalization type.

  • beta:

    The descriptor of the beta.

  • gamma:

    The descriptor of the gamma.

  • momentum:

    A value, between 0 and 1, the normalization operation uses to update the moving mean and moving variance during training.

  • epsilon:

    The epsilon in the computation of the standard deviation.

  • activation:

    The activation function that the layer applies to the output.