Struct dfdx::nn::modules::LayerNorm1D
source · pub struct LayerNorm1D<const M: usize, E: Dtype, D: DeviceStorage> {
pub gamma: Tensor<Rank1<M>, E, D>,
pub beta: Tensor<Rank1<M>, E, D>,
pub epsilon: E,
}
Expand description
Implements layer normalization as described in Layer Normalization.
This calls normalize() on the last axis of the input to normalize to 0 mean and unit std dev, and then does an element-wise affine transform using learnable parameters Self::gamma and Self::beta.
Self::epsilon is passed to normalize() and added to the variance to ensure big enough numbers. It defaults to 1e-5
.
Generics
M
The size of the affine transform tensors.
Examples
type Model = LayerNorm1D<5>;
let model = dev.build_module::<Model, f32>();
let _: Tensor<Rank1<5>, f32, _> = model.forward(dev.zeros::<Rank1<5>>());
Fields§
§gamma: Tensor<Rank1<M>, E, D>
§beta: Tensor<Rank1<M>, E, D>
§epsilon: E
Trait Implementations§
source§impl<const M: usize, E: Clone + Dtype, D: Clone + DeviceStorage> Clone for LayerNorm1D<M, E, D>
impl<const M: usize, E: Clone + Dtype, D: Clone + DeviceStorage> Clone for LayerNorm1D<M, E, D>
source§fn clone(&self) -> LayerNorm1D<M, E, D>
fn clone(&self) -> LayerNorm1D<M, E, D>
Returns a copy of the value. Read more
1.0.0 · source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
Performs copy-assignment from
source
. Read moresource§impl<const M: usize, E: Debug + Dtype, D: Debug + DeviceStorage> Debug for LayerNorm1D<M, E, D>
impl<const M: usize, E: Debug + Dtype, D: Debug + DeviceStorage> Debug for LayerNorm1D<M, E, D>
source§impl<B: Dim, const M: usize, E: Dtype, D: Device<E>, T: Tape<E, D>> Module<Tensor<(B, Const<M>), E, D, T>> for LayerNorm1D<M, E, D>
impl<B: Dim, const M: usize, E: Dtype, D: Device<E>, T: Tape<E, D>> Module<Tensor<(B, Const<M>), E, D, T>> for LayerNorm1D<M, E, D>
source§impl<B: Dim, S: Dim, const M: usize, E: Dtype, D: Device<E>, T: Tape<E, D>> Module<Tensor<(B, S, Const<M>), E, D, T>> for LayerNorm1D<M, E, D>
impl<B: Dim, S: Dim, const M: usize, E: Dtype, D: Device<E>, T: Tape<E, D>> Module<Tensor<(B, S, Const<M>), E, D, T>> for LayerNorm1D<M, E, D>
source§impl<const M: usize, E: Dtype, D: Device<E>, T: Tape<E, D>> Module<Tensor<(Const<M>,), E, D, T>> for LayerNorm1D<M, E, D>
impl<const M: usize, E: Dtype, D: Device<E>, T: Tape<E, D>> Module<Tensor<(Const<M>,), E, D, T>> for LayerNorm1D<M, E, D>
source§impl<const M: usize, E: Dtype, D: Device<E>> TensorCollection<E, D> for LayerNorm1D<M, E, D>
impl<const M: usize, E: Dtype, D: Device<E>> TensorCollection<E, D> for LayerNorm1D<M, E, D>
§type To<E2: Dtype, D2: Device<E2>> = LayerNorm1D<M, E2, D2>
type To<E2: Dtype, D2: Device<E2>> = LayerNorm1D<M, E2, D2>
Type alias that specifies the how a module’s type changes when using a different dtype and/or
device.
source§fn iter_tensors<V: ModuleVisitor<Self, E, D>>(
visitor: &mut V
) -> Result<Option<Self::To<V::E2, V::D2>>, V::Err>
fn iter_tensors<V: ModuleVisitor<Self, E, D>>( visitor: &mut V ) -> Result<Option<Self::To<V::E2, V::D2>>, V::Err>
Specifies how to iterate through tensors or modules containted within this module, and how
to contruct this module given values for its fields. Returns
Err(_)
to indicate an error,
Ok(None)
to indicate that there is no error and a module has not been built, and
Ok(Some(_))
contains Self::Output<E2, D2>
source§fn module<F1, F2, Field>(
name: &str,
get_ref: F1,
get_mut: F2
) -> ModuleField<'_, F1, F2, Self, Field>where
F1: FnMut(&Self) -> &Field,
F2: FnMut(&mut Self) -> &mut Field,
Field: TensorCollection<E, D>,
fn module<F1, F2, Field>( name: &str, get_ref: F1, get_mut: F2 ) -> ModuleField<'_, F1, F2, Self, Field>where F1: FnMut(&Self) -> &Field, F2: FnMut(&mut Self) -> &mut Field, Field: TensorCollection<E, D>,
Creates a ModuleFields that represents a field that may contain one or more tensors. Read more
source§fn tensor<F1, F2, S>(
name: &str,
get_ref: F1,
get_mut: F2,
options: TensorOptions<S, E, D>
) -> TensorField<'_, F1, F2, Self, S, E, D>where
F1: FnMut(&Self) -> &Tensor<S, E, D>,
F2: FnMut(&mut Self) -> &mut Tensor<S, E, D>,
S: Shape,
fn tensor<F1, F2, S>( name: &str, get_ref: F1, get_mut: F2, options: TensorOptions<S, E, D> ) -> TensorField<'_, F1, F2, Self, S, E, D>where F1: FnMut(&Self) -> &Tensor<S, E, D>, F2: FnMut(&mut Self) -> &mut Tensor<S, E, D>, S: Shape,
Creates a ModuleFields that represents a tensor field. Read more