Struct Flatten 

pub struct Flatten<D: Dtype, IT, OT> { /* private fields */ }

Flatten layer: collapses all spatial dimensions into a single feature vector.

Input shape: [N, C, H, W] (rank 4) Output shape: [N, C * H * W] (rank 2)

Type parameters:

• D — element dtype
• IT — input tensor type (rank 4: [N, C, H, W])
• OT — output tensor type (rank 2: [N, C*H*W])

Tensor bounds are on impls, not the struct, so SymTensor can have its own Layer impl without a coherence conflict.

Implementations

impl<D: Dtype, IT, OT> Flatten<D, IT, OT>

pub fn new() -> Self

Creates a new Flatten layer.

Trait Implementations

impl<D: Dtype, IT, OT> Default for Flatten<D, IT, OT>

fn default() -> Self

Returns the “default value” for a type.

impl<D: Dtype, IT: Tensor<D, 4> + EagerTensor, OT: Tensor<D, 2>> Layer<IT> for Flatten<D, IT, OT>

type Output = OT

This layer’s output type.

fn call(&self, _input: IT) -> Self::Output

Applies this layer to input.

impl<D: Dtype> Layer<SymTensor> for Flatten<D, SymTensor, SymTensor>

type Output = SymTensor

This layer’s output type.

fn call(&self, input: SymTensor) -> SymTensor

Applies this layer to input.