Enum Padding 

#[non_exhaustive]
pub enum Padding {
    Explicit(Vec<i64>),
    Same,
    Valid,
}

Convolution / pooling padding specification.

Upstream NIR accepts integer extents or the string modes "same" / "valid".

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

Explicit(Vec<i64>)

Explicit per-axis padding extents (length 1 for 1d, 2 for 2d, …).

Same

Pad so that spatial output size matches input ("same").

Valid

No padding ("valid").

Implementations

impl Padding

pub fn single(value: i64) -> Self

Single-axis explicit padding.

pub fn pair(h: i64, w: i64) -> Self

Two-axis explicit padding (h, w).

Trait Implementations

impl Clone for Padding

fn clone(&self) -> Padding

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Padding

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

Formats the value using the given formatter.

impl PartialEq for Padding

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for Padding