Struct dfdx::nn::modules::Conv2D

pub struct Conv2D<const IN_CHAN: usize, const OUT_CHAN: usize, const KERNEL_SIZE: usize, const STRIDE: usize, const PADDING: usize, const DILATION: usize, const GROUPS: usize, E: Dtype, D: Storage<E>>where
    Const<{ _ }>: Sized,{
    pub weight: Tensor<Rank4<OUT_CHAN, { _ }, KERNEL_SIZE, KERNEL_SIZE>, E, D>,
}

Requires Nightly Performs unbiased 2d convolutions on 3d and 4d images.

Pytorch Equivalent: torch.nn.Conv2d(..., bias=False)

To create a biased conv, combine with crate::nn::modules::Bias2D:

type BiasedConv = (Conv2D<3, 5, 4>, Bias2D<5>);

Generics:

• IN_CHAN: The number of input channels in an image.
• OUT_CHAN: The number of channels in the output of the layer.
• KERNEL_SIZE: The size of the kernel applied to both width and height of the images.
• STRIDE: How far to move the kernel each step. Defaults to 1
• PADDING: How much zero padding to add around the images. Defaults to 0.
• DILATION: Controls the spacing between kernel points. Defaults to 1.
• GROUPS: Controls the connections between inputs and outputs. IN_CHAN and OUT_CHAN must both be divisible by GROUPS. For example,

See conv animations for helpful visualization of all of these parameters.

Fields

weight: Tensor<Rank4<OUT_CHAN, { _ }, KERNEL_SIZE, KERNEL_SIZE>, E, D>

Trait Implementations

impl<const IN_CHAN: usize, const OUT_CHAN: usize, const KERNEL_SIZE: usize, const STRIDE: usize, const PADDING: usize, const DILATION: usize, const GROUPS: usize, E: Clone + Dtype, D: Clone + Storage<E>> Clone for Conv2D<IN_CHAN, OUT_CHAN, KERNEL_SIZE, STRIDE, PADDING, DILATION, GROUPS, E, D>where Const<{ _ }>: Sized,

fn clone( &self ) -> Conv2D<IN_CHAN, OUT_CHAN, KERNEL_SIZE, STRIDE, PADDING, DILATION, GROUPS, E, D>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<const IN_CHAN: usize, const OUT_CHAN: usize, const KERNEL_SIZE: usize, const STRIDE: usize, const PADDING: usize, const DILATION: usize, const GROUPS: usize, E: Debug + Dtype, D: Debug + Storage<E>> Debug for Conv2D<IN_CHAN, OUT_CHAN, KERNEL_SIZE, STRIDE, PADDING, DILATION, GROUPS, E, D>where Const<{ _ }>: Sized,

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

Formats the value using the given formatter.

impl<const I: usize, const O: usize, const K: usize, const S: usize, const P: usize, const L: usize, const G: usize, E, D, Img> Module<Img> for Conv2D<I, O, K, S, P, L, G, E, D>where Const<{ _ }>: Sized, E: Dtype, D: Device<E>, (Img, Tensor<Rank4<O, { _ }, K, K>, E, D>): TryConv2D<Const<S>, Const<P>, Const<L>, Const<G>>,

type Output = <(Img, Tensor<(Const<O>, Const<{ I / G }>, Const<K>, Const<K>), E, D, NoneTape>) as TryConv2D<Const<S>, Const<P>, Const<L>, Const<G>>>::Convolved

The type that this unit produces given Input.

type Error = <(Img, Tensor<(Const<O>, Const<{ I / G }>, Const<K>, Const<K>), E, D, NoneTape>) as TryConv2D<Const<S>, Const<P>, Const<L>, Const<G>>>::Error

fn try_forward(&self, x: Img) -> Result<Self::Output, Self::Error>

fn forward(&self, input: Input) -> Self::Output

Forward Input through the module and produce Module::Output.

impl<const I: usize, const O: usize, const K: usize, const S: usize, const P: usize, const L: usize, const G: usize, E, D> TensorCollection<E, D> for Conv2D<I, O, K, S, P, L, G, E, D>where Const<{ _ }>: Sized, E: Dtype + Float + SampleUniform, D: Device<E>,

type To<E2: Dtype, D2: Device<E2>> = Conv2D<I, O, K, S, P, L, G, E2, D2>

Type alias that specifies the how a module’s type changes when using a different dtype and/or device.

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>

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.

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.

fn scalar<F1, F2, N>( name: &str, get_ref: F1, get_mut: F2, options: ScalarOptions<N> ) -> ScalarField<'_, F1, F2, Self, N>where F1: FnMut(&Self) -> &N, F2: FnMut(&mut Self) -> &mut N, N: NumCast,

Creates a ModuleFields that represents a scalar field.

impl<const I: usize, const O: usize, const K: usize, const S: usize, const P: usize, const L: usize, const G: usize, E: Dtype, D: Storage<E>> NonMutableModule for Conv2D<I, O, K, S, P, L, G, E, D>where Const<{ _ }>: Sized,