Struct Conv3d 

pub struct Conv3d(/* private fields */);

Named-channel 3D cross-correlation followed by bias addition.

The three spatial axes conventionally represent depth, height, and width; kernel, stride, and symmetric padding arrays use that same supplied order. Stride defaults to [1, 1, 1], padding to [0, 0, 0], and groups to 1. Dilation and asymmetric padding are not supported.

The backend computes patch indices from compact geometry and uses a deterministic input-centric derivative, but materializes the FP32 patch tensor before contraction. Downstream layout operations still have the generic index-plan limits documented by Axis.

Implementations

impl Conv3d

pub fn new( input: Axis, output: Dim, spatial: [Axis; 3], kernel: [usize; 3], ) -> Self

pub fn stride(self, stride: [usize; 3]) -> Self

Set stride in the order of the supplied spatial axes.

pub fn padding(self, padding: [usize; 3]) -> Self

Set symmetric zero-padding in the order of the supplied spatial axes.

pub fn groups(self, groups: usize) -> Self

Partition input and output channels into independent convolution groups.

pub fn bias(self, bias: bool) -> Self

Disable the learned bias term, matching crate::Linear::bias. Has no effect once build has already allocated parameters; call it before build.

Trait Implementations

impl Module for Conv3d

fn output_shape(&self, input: &Shape) -> Result<Shape>

fn build(&mut self, input: &Shape, device: &Device, seed: u64) -> Result<Shape>

Infer extents, validate contracts, allocate and initialize parameters; never run forward. Rebuilding with compatible extents/device preserves parameters and ignores the new seed.

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

Evaluation semantics: no mode flag on self, no randomness beyond what forward itself already defines. This never changes with the arrival of forward_training.

fn named_parameters(&self) -> Vec<(String, Parameter)>

Structural paths identify slots; ParamId identifies shared storage within this process.

fn forward_training( &self, input: &Tensor, pass: &mut TrainingPass, ) -> Result<Tensor>

Training-mode forward pass, driven by an explicit TrainingPass rather than a flag on self. The default implementation ignores pass and calls Module::forward, which is exactly today’s behavior for every module without a train/eval distinction. Override this (never forward, which must stay evaluation semantics) to add explicit randomness or a train/eval difference, such as Dropout. A container that holds child modules (Sequential, or any future one) overrides this to thread the same pass through every child in order, so their draws stay distinct within one step.

fn parameter(&self, name: &str) -> Result<Parameter>

fn parameters(&self) -> Vec<Parameter>

fn zero_grad(&self)

fn named_states(&self) -> Vec<(String, State)>

Persistent non-parameter tensors this module owns across training steps – BatchNorm-style running statistics. Same slot-path convention as Module::named_parameters: a container (Sequential or any future one) aggregates every child’s own states under "{index}.{name}". Defaults empty, exactly like named_parameters, since most modules have no persistent state. A State here is never written directly; only TrainingPass::commit writes one, after staging in forward_training.

fn state(&self, name: &str) -> Result<State>

fn states(&self) -> Vec<State>