Contents

init(normalizedShape:beta:gamma:varianceEpsilon:)

Creates a normalization layer with a shape, beta and gamma tensors, and variance epsilon you specify.

Declaration

convenience init?(normalizedShape: [Int], beta: MLCTensor, gamma: MLCTensor, varianceEpsilon: Float)

Parameters

  • normalizedShape:

    The shape of the axes where normalization occurs.

  • beta:

    The beta tensor.

  • gamma:

    The gamma tensor.

  • varianceEpsilon:

    The variance epsilon you use for numerical stability.

See Also

Creating Layer Normalization Layers