Struct syntaxdot_transformers::layers::LayerNorm

pub struct LayerNorm { /* private fields */ }

Layer that applies layer normalization.

Implementations

impl LayerNorm

pub fn new<'a>(
    vs: impl Borrow<PathExt<'a>>,
    normalized_shape: impl Into<Vec<i64>>,
    eps: f64,
    elementwise_affine: bool
) -> Self

Construct a layer normalization layer.

The mean and standard deviation are computed over the last number of dimensions with the shape defined by normalized_shape. If elementwise_affine is True, a learnable affine transformation of the shape normalized_shape is added after normalization.

Trait Implementations

impl Debug for LayerNorm

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl FallibleModule for LayerNorm

type Error = TransformerError

The error type.

fn forward(&self, input: &Tensor) -> Result<Tensor, Self::Error>

Apply the module.