Struct Conv2d 

pub struct Conv2d {
    pub weight: Tensor,
    pub stride: Vec<i64>,
    pub padding: Padding,
    pub dilation: Vec<i64>,
    pub groups: i64,
    pub bias: Tensor,
    pub input_shape: Option<Vec<usize>>,
    pub metadata: MetadataMap,
}

2-D convolution.

Fields

weight: Tensor

Kernel weights, typically (C_out, C_in, Kh, Kw).

stride: Vec<i64>

Stride per spatial axis (or single value expanded later).

padding: Padding

Padding specification.

dilation: Vec<i64>

Dilation per spatial axis.

groups: i64

Grouped convolution groups.

bias: Tensor

Bias of shape (C_out,) (required on the NIR wire).

input_shape: Option<Vec<usize>>

Optional spatial input (N_x, N_y).

metadata: MetadataMap

Free-form node metadata (Python metadata dict).

Trait Implementations

impl Clone for Conv2d

fn clone(&self) -> Conv2d

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Conv2d

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

Formats the value using the given formatter.

impl PartialEq for Conv2d

fn eq(&self, other: &Conv2d) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for Conv2d