Struct nncombinator::layer::LinearLayer

source · [−]

pub struct LinearLayer<U, C, P, D, I, const NI: usize, const NO: usize>where
    P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain<U> + Loss<U>,
    U: Default + Clone + Copy + Send + UnitValue<U>,
    D: Device<U>,
    I: Debug + Send + Sync,{ /* private fields */ }

Expand description

Linear Layer Implementation

Implementations

source

impl<U, P, I, const NI: usize, const NO: usize> LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

source

pub fn new<UI: FnMut() -> U, BI: FnMut() -> U>(
 parent: P,
 device: &DeviceCpu,
 ui: UI,
 bi: BI
) -> LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>

Create and return an instance of LinearLayer

Arguments

parent - upper layer
device - Device object used for neural network computation
ui - Callback to generate weight of unit
bi - Callback to generate weight of bias

source

impl<U, P, I, const NI: usize, const NO: usize> LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,
 DeviceGpu: Device,

source

pub fn new<UI: FnMut() -> U, BI: FnMut() -> U>(
 parent: P,
 device: &DeviceGpu,
 ui: UI,
 bi: BI
) -> Result<LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>, CudaError>

Attempt to create and return an instance of LinearLayer.

Arguments

parent - upper layer
device - Device object used for neural network computation
ui - Callback to generate weight of unit
bi - Callback to generate weight of bias

Errors

This function may return the following errors

CudaError

Trait Implementations

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> AskDiffInput for LinearLayer<U, C, P, D, I, NI, NO>where
 P: PreTrain<U, OutStack = <<Self as PreTrain>::OutStack as Stack>::Remaining> + ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + Loss + AskDiffInput,
 U: Default + Clone + Copy + Send + UnitValue,
 D: Device,
 I: Debug + Send + Sync,
 Self: PreTrain,

type DiffInput = >::DiffInput

Diff Input to this layer of the neural network

source

fn ask_diff_input(&self, stack: &Self::OutStack) -> Self::DiffInput

Data inquiry for creating difference information Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> Backward<U, &Arr<U, NO>, Result<Arr<U, NI>, TrainingError>> for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss,
 U: Default + Clone + Copy + UnitValue,
 D: Device + DeviceLinear<U, C, NI, NO>,
 I: Debug + Send + Sync,

source

fn backward(&mut self, input: &Arr<U, NO>) -> Result<Arr<U, NI>, TrainingError>

Back propagation of errors #Arguments Read more

source

impl<U, P, I, const NI: usize, const NO: usize> BackwardAll for LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 P: BackwardAll<U, LossInput = Arr<U, NI>> + ForwardAll<Input = I, Output = Arr<U, NI>> + PreTrain + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

type LossInput = Arr<U, NO>

Losses during neural network training

source

fn backward_all<OP: Optimizer, L: LossFunction>(
 &mut self,
 input: Self::LossInput,
 stack: Self::OutStack,
 optimizer: &mut OP,
 lossf: &L
) -> Result<(), TrainingError>

Back propagation of errors #Arguments Read more

source

fn is_canonical_link<L: LossFunction>(&self, _: &L) -> bool

source

impl<U, P, I, const NI: usize, const NO: usize> BackwardAll for LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 P: BackwardAll<U, LossInput = Arr<U, NI>> + ForwardAll<Input = I, Output = Arr<U, NI>> + PreTrain + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,
 DeviceGpu: Device + DeviceLinear<U, CachedTensor<U, Arr2<U, NI, NO>>, NI, NO>,

type LossInput = Arr<U, NO>

Losses during neural network training

source

fn backward_all<OP: Optimizer, L: LossFunction>(
 &mut self,
 input: Self::LossInput,
 stack: Self::OutStack,
 optimizer: &mut OP,
 lossf: &L
) -> Result<(), TrainingError>

Back propagation of errors #Arguments Read more

source

fn is_canonical_link<L: LossFunction>(&self, _: &L) -> bool

source

impl<U, P, I, const NI: usize, const NO: usize> BatchBackward for LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward + BatchPreTrainBase + BatchPreTrain + BatchBackward + BatchLoss<U, BatchLossInput = VecArr<U, Arr<U, NI>>>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

type BatchLossInput = VecArr<U, Arr<U, NO>>

Losses during neural network training for batch execution

source

fn batch_backward<OP: Optimizer, L: LossFunction>(
 &mut self,
 input: Self::BatchLossInput,
 stack: Self::BatchOutStack,
 optimizer: &mut OP,
 lossf: &L
) -> Result<(), TrainingError>

Back propagation of errors #Arguments Read more

source

impl<U, P, I, const NI: usize, const NO: usize> BatchBackward for LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward + BatchPreTrainBase + BatchPreTrain + BatchBackward + BatchLoss<U, BatchLossInput = VecArr<U, Arr<U, NI>>>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,
 DeviceGpu: Device + DeviceLinear<U, CachedTensor<U, Arr2<U, NI, NO>>, NI, NO>,

type BatchLossInput = VecArr<U, Arr<U, NO>>

Losses during neural network training for batch execution

source

fn batch_backward<OP: Optimizer, L: LossFunction>(
 &mut self,
 input: Self::BatchLossInput,
 stack: Self::BatchOutStack,
 optimizer: &mut OP,
 lossf: &L
) -> Result<(), TrainingError>

Back propagation of errors #Arguments Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> BatchForward for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward,
 D: Device + DeviceLinear<U, C, NI, NO>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

source

fn batch_forward(
&self,
input: Self::BatchInput
) -> Result<Self::BatchOutput, TrainingError>

Forward propagation #Arguments Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> BatchForwardBase for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>>,
 U: Default + Clone + Copy + Send + UnitValue,
 D: Device,
 I: Debug + Send + Sync,
 Self: ForwardAll,

type BatchInput = VecArr<U, I>

Input to this layer of the neural network for batch execution

type BatchOutput = VecArr<U, Arr<U, NO>>

Output from this layer of the neural network for batch execution

source

impl<U, P, I, const NI: usize, const NO: usize> BatchLoss for LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward + BatchPreTrainBase + BatchPreTrain + BatchBackward + BatchLoss<U, BatchLossInput = VecArr<U, Arr<U, NI>>>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

source

fn batch_loss<L: LossFunction>(
 &self,
 loss: Self::BatchLossInput,
 _lossf: &L,
 stack: Self::BatchOutStack
) -> Result<(Self::BatchOutStack, Self::BatchLossInput), TrainingError>

Error Calculation Read more

source

impl<U, P, I, const NI: usize, const NO: usize> BatchLoss for LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward + BatchPreTrainBase + BatchPreTrain + BatchBackward + BatchLoss<U, BatchLossInput = VecArr<U, Arr<U, NI>>>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,
 DeviceGpu: Device,
 Self: Loss + BatchBackward,

source

fn batch_loss<L: LossFunction>(
 &self,
 loss: Self::BatchLossInput,
 _lossf: &L,
 stack: Self::BatchOutStack
) -> Result<(Self::BatchOutStack, Self::BatchLossInput), TrainingError>

Error Calculation Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> BatchPreTrain for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward + BatchPreTrainBase + BatchPreTrain,
 U: Default + Clone + Copy + Send + UnitValue,
 D: Device + DeviceLinear<U, C, NI, NO>,
 I: Debug + Send + Sync,

source

fn batch_pre_train(
&self,
input: Self::BatchInput
) -> Result<Self::BatchOutStack, TrainingError>

Perform forward propagation required to perform error back propagation #Arguments Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> BatchPreTrainBase for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + BatchForwardBase<BatchInput = VecArr<U, I>, BatchOutput = VecArr<U, Arr<U, NI>>> + BatchForward + BatchPreTrainBase,
 U: Default + Clone + Copy + Send + UnitValue,
 D: Device,
 I: Debug + Send + Sync,
 Self: PreTrain,

type BatchOutStack = Cons<>::BatchOutStack, <LinearLayer<U, C, P, D, I, NI, NO> as BatchForwardBase>::BatchOutput>

Type of object to keep the results of forward propagation needed to perform error back propagation for batch execution. Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> Forward<Arr<U, NI>, Result<Arr<U, NO>, EvaluateError>> for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 D: Device + DeviceLinear<U, C, NI, NO>,
 I: Debug + Send + Sync,

source

fn forward(&self, input: &Arr<U, NI>) -> Result<Arr<U, NO>, EvaluateError>

Forward propagation implementation Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> ForwardAll for LinearLayer<U, C, P, D, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss,
 D: Device + DeviceLinear<U, C, NI, NO>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

type Input = I

Input to this layer of the neural network

type Output = Arr<U, NO>

Output from this layer of the neural network

source

fn forward_all(&self, input: Self::Input) -> Result<Self::Output, EvaluateError>

Forward propagation #Arguments Read more

source

impl<U, P, I, const NI: usize, const NO: usize> Loss for LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 P: PreTrain + ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

source

fn loss<L: LossFunction>(
 &mut self,
 loss: Self::LossInput,
 _lossf: &L,
 stack: Self::OutStack
) -> Result<(Self::OutStack, Self::LossInput), TrainingError>

Error Calculation Read more

source

impl<U, P, I, const NI: usize, const NO: usize> Loss for LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 P: PreTrain + ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + Loss,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,
 DeviceGpu: Device,
 Self: BackwardAll,

source

fn loss<L: LossFunction>(
 &mut self,
 loss: Self::LossInput,
 _lossf: &L,
 stack: Self::OutStack
) -> Result<(Self::OutStack, Self::LossInput), TrainingError>

Error Calculation Read more

source

impl<T, U, P, I, const NI: usize, const NO: usize> Persistence<U, T, Linear> for LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 T: LinearPersistence,
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + Persistence<U, T, Linear>,
 U: Default + Clone + Copy + UnitValue,
 I: Debug + Send + Sync,

source

fn load(&mut self, persistence: &mut T) -> Result<(), ConfigReadError>

Load Model Read more

source

fn save(&mut self, persistence: &mut T) -> Result<(), PersistenceError>

Save Model Read more

source

impl<T, U, P, I, const NI: usize, const NO: usize> Persistence<U, T, Linear> for LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 T: LinearPersistence,
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + Persistence<U, T, Linear>,
 U: Default + Clone + Copy + UnitValue,
 I: Debug + Send + Sync,
 DeviceGpu: Device,

source

fn load(&mut self, persistence: &mut T) -> Result<(), ConfigReadError>

Load Model Read more

source

fn save(&mut self, persistence: &mut T) -> Result<(), PersistenceError>

Save Model Read more

source

impl<U, P, I, const NI: usize, const NO: usize> Persistence<U, TextFilePersistence, Specialized> for LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + Persistence<U, TextFilePersistence, Specialized>,
 U: Default + Clone + Copy + UnitValue + FromStr,
 I: Debug + Send + Sync,
 ConfigReadError: From<::Err>,

source

fn load(
&mut self,
persistence: &mut TextFilePersistence
) -> Result<(), ConfigReadError>

Load Model Read more

source

fn save(
&mut self,
persistence: &mut TextFilePersistence
) -> Result<(), PersistenceError>

Save Model Read more

source

impl<U, P, I, const NI: usize, const NO: usize> Persistence<U, TextFilePersistence, Specialized> for LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu, I, NI, NO>where
 P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain + Loss + Persistence<U, TextFilePersistence, Specialized>,
 U: Default + Clone + Copy + UnitValue + FromStr,
 I: Debug + Send + Sync,
 DeviceGpu: Device,
 ConfigReadError: From<::Err>,

source

fn load(
&mut self,
persistence: &mut TextFilePersistence
) -> Result<(), ConfigReadError>

Load Model Read more

source

fn save(
&mut self,
persistence: &mut TextFilePersistence
) -> Result<(), PersistenceError>

Save Model Read more

source

impl<U, C, P, D, I, const NI: usize, const NO: usize> PreTrain for LinearLayer<U, C, P, D, I, NI, NO>where
 P: PreTrain + ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + Loss,
 D: Device + DeviceLinear<U, C, NI, NO>,
 U: Default + Clone + Copy + Send + UnitValue,
 I: Debug + Send + Sync,

type OutStack = Cons<>::OutStack, <LinearLayer<U, C, P, D, I, NI, NO> as ForwardAll>::Output>

Type of object to keep the results of forward propagation needed to perform error back propagation.

source

fn pre_train(&self, input: Self::Input) -> Result<Self::OutStack, EvaluateError>

Perform forward propagation required to perform error back propagation Read more

Auto Trait Implementations

impl<U, C, P, D, I, const NI: usize, const NO: usize> RefUnwindSafe for LinearLayer<U, C, P, D, I, NI, NO>where
 C: RefUnwindSafe,
 D: RefUnwindSafe,
 P: RefUnwindSafe,
 U: RefUnwindSafe,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Send for LinearLayer<U, C, P, D, I, NI, NO>where
 C: Send,
 D: Send,
 P: Send,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Sync for LinearLayer<U, C, P, D, I, NI, NO>where
 C: Sync,
 D: Sync,
 P: Sync,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Unpin for LinearLayer<U, C, P, D, I, NI, NO>where
 C: Unpin,
 D: Unpin,
 P: Unpin,

impl<U, C, P, D, I, const NI: usize, const NO: usize> UnwindSafe for LinearLayer<U, C, P, D, I, NI, NO>where
 C: UnwindSafe,
 D: UnwindSafe,
 P: UnwindSafe,
 U: UnwindSafe,

Blanket Implementations

source

impl<T> AddLayer for Twhere
T: ForwardAll,

source

fn add_layer<C, F>(self, f: F) -> Cwhere
C: ForwardAll,
F: FnOnce(T) -> C,

Adding Layers Read more

source

impl<T, U> AddLayerTrain for Twhere
T: PreTrain,
U: UnitValue,

source

fn add_layer_train<C, F>(self, f: F) -> Cwhere
C: Train,
F: FnOnce(T) -> C,

Adding Layers Read more

source

impl<T> Any for Twhere
T: 'static + ?Sized,

source

fn type_id(&self) -> TypeId

Gets the TypeId of self. Read more

source

impl<T> Borrow<T> for Twhere
T: ?Sized,

const: unstable · source

fn borrow(&self) -> &T

Immutably borrows from an owned value. Read more

source

impl<T> BorrowMut<T> for Twhere
T: ?Sized,

const: unstable · source

fn borrow_mut(&mut self) -> &mut T

Mutably borrows from an owned value. Read more

source

impl<T, U> ForwardDiff for Twhere
T: PreTrain,
U: UnitValue,

source

fn forward_diff(
&self,
input: <T as ForwardAll>::Input
) -> Result<<T as PreTrain>::OutStack, EvaluateError>

Forward propagation (differential application) Read more

source

impl<T> From<T> for T

const: unstable · source

fn from(t: T) -> T

Returns the argument unchanged.

source

impl<T, U> Into for Twhere
U: From<T>,

const: unstable · source

fn into(self) -> U

Calls U::from(self).

That is, this conversion is whatever the implementation of From<T> for U chooses to do.

impl<T> Pointable for T

const ALIGN: usize = mem::align_of::<T>()

The alignment of pointer.

type Init = T

The type for initializers.

unsafe fn init(init: <T as Pointable>::Init) -> usize

Initializes a with the given initializer. Read more

unsafe fn deref<'a>(ptr: usize) -> &'a T

Dereferences the given pointer. Read more

unsafe fn deref_mut<'a>(ptr: usize) -> &'a mut T

Mutably dereferences the given pointer. Read more

unsafe fn drop(ptr: usize)

Drops the object pointed to by the given pointer. Read more

source

impl<T> TryAddLayer for Twhere
T: ForwardAll,

source

fn try_add_layer<C, F>(self, f: F) -> Result<C, DeviceError>where
C: ForwardAll,
F: FnOnce(T) -> Result<C, DeviceError>,

Adding Layers Read more

source

impl<T, U> TryAddLayerTrain for Twhere
T: PreTrain,
U: UnitValue,

source

fn try_add_layer_train<C, F>(self, f: F) -> Result<C, DeviceError>where
C: Train,
F: FnOnce(T) -> Result<C, DeviceError>,

Adding Layers Read more

source

impl<T, U> TryFrom for Twhere
U: Into<T>,

type Error = Infallible

The type returned in the event of a conversion error.

const: unstable · source

fn try_from(value: U) -> Result<T, <T as TryFrom>::Error>

Performs the conversion.

source

impl<T, U> TryInto for Twhere
U: TryFrom<T>,

type Error = >::Error

The type returned in the event of a conversion error.

const: unstable · source

fn try_into(self) -> Result<U, >::Error>

Performs the conversion.

Struct nncombinator::layer::LinearLayer

Implementations

impl<U, P, I, const NI: usize, const NO: usize> LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu<U>, I, NI, NO>where P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain<U> + Loss<U>, U: Default + Clone + Copy + Send + UnitValue<U>, I: Debug + Send + Sync,

pub fn new<UI: FnMut() -> U, BI: FnMut() -> U>( parent: P, device: &DeviceCpu<U>, ui: UI, bi: BI) -> LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu<U>, I, NI, NO>

pub fn new<UI: FnMut() -> U, BI: FnMut() -> U>( parent: P, device: &DeviceGpu<U>, ui: UI, bi: BI) -> Result<LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu<U>, I, NI, NO>, CudaError>

Trait Implementations

type DiffInput = <P as AskDiffInput<U>>::DiffInput

fn ask_diff_input(&self, stack: &Self::OutStack) -> Self::DiffInput

fn backward(&mut self, input: &Arr<U, NO>) -> Result<Arr<U, NI>, TrainingError>

type LossInput = Arr<U, NO>

fn backward_all<OP: Optimizer<U>, L: LossFunction<U>>( &mut self, input: Self::LossInput, stack: Self::OutStack, optimizer: &mut OP, lossf: &L) -> Result<(), TrainingError>

fn is_canonical_link<L: LossFunction<U>>(&self, _: &L) -> bool

type LossInput = Arr<U, NO>

fn backward_all<OP: Optimizer<U>, L: LossFunction<U>>( &mut self, input: Self::LossInput, stack: Self::OutStack, optimizer: &mut OP, lossf: &L) -> Result<(), TrainingError>

fn is_canonical_link<L: LossFunction<U>>(&self, _: &L) -> bool

type BatchLossInput = VecArr<U, Arr<U, NO>>

fn batch_backward<OP: Optimizer<U>, L: LossFunction<U>>( &mut self, input: Self::BatchLossInput, stack: Self::BatchOutStack, optimizer: &mut OP, lossf: &L) -> Result<(), TrainingError>

type BatchLossInput = VecArr<U, Arr<U, NO>>

fn batch_backward<OP: Optimizer<U>, L: LossFunction<U>>( &mut self, input: Self::BatchLossInput, stack: Self::BatchOutStack, optimizer: &mut OP, lossf: &L) -> Result<(), TrainingError>

fn batch_forward( &self, input: Self::BatchInput) -> Result<Self::BatchOutput, TrainingError>

type BatchInput = VecArr<U, I>

type BatchOutput = VecArr<U, Arr<U, NO>>

fn batch_loss<L: LossFunction<U>>( &self, loss: Self::BatchLossInput, _lossf: &L, stack: Self::BatchOutStack) -> Result<(Self::BatchOutStack, Self::BatchLossInput), TrainingError>

fn batch_loss<L: LossFunction<U>>( &self, loss: Self::BatchLossInput, _lossf: &L, stack: Self::BatchOutStack) -> Result<(Self::BatchOutStack, Self::BatchLossInput), TrainingError>

fn batch_pre_train( &self, input: Self::BatchInput) -> Result<Self::BatchOutStack, TrainingError>

type BatchOutStack = Cons<<P as BatchPreTrainBase<U>>::BatchOutStack, <LinearLayer<U, C, P, D, I, NI, NO> as BatchForwardBase>::BatchOutput>

fn forward(&self, input: &Arr<U, NI>) -> Result<Arr<U, NO>, EvaluateError>

type Input = I

type Output = Arr<U, NO>

fn forward_all(&self, input: Self::Input) -> Result<Self::Output, EvaluateError>

fn loss<L: LossFunction<U>>( &mut self, loss: Self::LossInput, _lossf: &L, stack: Self::OutStack) -> Result<(Self::OutStack, Self::LossInput), TrainingError>

fn loss<L: LossFunction<U>>( &mut self, loss: Self::LossInput, _lossf: &L, stack: Self::OutStack) -> Result<(Self::OutStack, Self::LossInput), TrainingError>

fn load(&mut self, persistence: &mut T) -> Result<(), ConfigReadError>

fn save(&mut self, persistence: &mut T) -> Result<(), PersistenceError>

fn load(&mut self, persistence: &mut T) -> Result<(), ConfigReadError>

fn save(&mut self, persistence: &mut T) -> Result<(), PersistenceError>

fn load( &mut self, persistence: &mut TextFilePersistence<U>) -> Result<(), ConfigReadError>

fn save( &mut self, persistence: &mut TextFilePersistence<U>) -> Result<(), PersistenceError>

fn load( &mut self, persistence: &mut TextFilePersistence<U>) -> Result<(), ConfigReadError>

fn save( &mut self, persistence: &mut TextFilePersistence<U>) -> Result<(), PersistenceError>

type OutStack = Cons<<P as PreTrain<U>>::OutStack, <LinearLayer<U, C, P, D, I, NI, NO> as ForwardAll>::Output>

fn pre_train(&self, input: Self::Input) -> Result<Self::OutStack, EvaluateError>

Auto Trait Implementations

impl<U, C, P, D, I, const NI: usize, const NO: usize> RefUnwindSafe for LinearLayer<U, C, P, D, I, NI, NO>where C: RefUnwindSafe, D: RefUnwindSafe, P: RefUnwindSafe, U: RefUnwindSafe,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Send for LinearLayer<U, C, P, D, I, NI, NO>where C: Send, D: Send, P: Send,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Sync for LinearLayer<U, C, P, D, I, NI, NO>where C: Sync, D: Sync, P: Sync,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Unpin for LinearLayer<U, C, P, D, I, NI, NO>where C: Unpin, D: Unpin, P: Unpin,

impl<U, C, P, D, I, const NI: usize, const NO: usize> UnwindSafe for LinearLayer<U, C, P, D, I, NI, NO>where C: UnwindSafe, D: UnwindSafe, P: UnwindSafe, U: UnwindSafe,

Blanket Implementations

impl<T> AddLayer for Twhere T: ForwardAll,

fn add_layer<C, F>(self, f: F) -> Cwhere C: ForwardAll, F: FnOnce(T) -> C,

impl<T, U> AddLayerTrain<U> for Twhere T: PreTrain<U>, U: UnitValue<U>,

fn add_layer_train<C, F>(self, f: F) -> Cwhere C: Train<U>, F: FnOnce(T) -> C,

impl<T> Any for Twhere T: 'static + ?Sized,

fn type_id(&self) -> TypeId

impl<T> Borrow<T> for Twhere T: ?Sized,

fn borrow(&self) -> &T

impl<T> BorrowMut<T> for Twhere T: ?Sized,

fn borrow_mut(&mut self) -> &mut T

impl<T, U> ForwardDiff<U> for Twhere T: PreTrain<U>, U: UnitValue<U>,

fn forward_diff( &self, input: <T as ForwardAll>::Input) -> Result<<T as PreTrain<U>>::OutStack, EvaluateError>

impl<T> From<T> for T

fn from(t: T) -> T

impl<T, U> Into<U> for Twhere U: From<T>,

fn into(self) -> U

impl<T> Pointable for T

const ALIGN: usize = mem::align_of::<T>()

type Init = T

unsafe fn init(init: <T as Pointable>::Init) -> usize

unsafe fn deref<'a>(ptr: usize) -> &'a T

unsafe fn deref_mut<'a>(ptr: usize) -> &'a mut T

unsafe fn drop(ptr: usize)

impl<T> TryAddLayer for Twhere T: ForwardAll,

fn try_add_layer<C, F>(self, f: F) -> Result<C, DeviceError>where C: ForwardAll, F: FnOnce(T) -> Result<C, DeviceError>,

impl<T, U> TryAddLayerTrain<U> for Twhere T: PreTrain<U>, U: UnitValue<U>,

fn try_add_layer_train<C, F>(self, f: F) -> Result<C, DeviceError>where C: Train<U>, F: FnOnce(T) -> Result<C, DeviceError>,

impl<T, U> TryFrom<U> for Twhere U: Into<T>,

type Error = Infallible

fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>

impl<T, U> TryInto<U> for Twhere U: TryFrom<T>,

impl<U, P, I, const NI: usize, const NO: usize> LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu<U>, I, NI, NO>where
P: ForwardAll<Input = I, Output = Arr<U, NI>> + BackwardAll<U, LossInput = Arr<U, NI>> + PreTrain<U> + Loss<U>,
U: Default + Clone + Copy + Send + UnitValue<U>,
I: Debug + Send + Sync,

pub fn new<UI: FnMut() -> U, BI: FnMut() -> U>(
parent: P,
device: &DeviceCpu<U>,
ui: UI,
bi: BI
) -> LinearLayer<U, Arr2<U, NI, NO>, P, DeviceCpu<U>, I, NI, NO>

pub fn new<UI: FnMut() -> U, BI: FnMut() -> U>(
parent: P,
device: &DeviceGpu<U>,
ui: UI,
bi: BI
) -> Result<LinearLayer<U, CachedTensor<U, Arr2<U, NI, NO>>, P, DeviceGpu<U>, I, NI, NO>, CudaError>

fn backward_all<OP: Optimizer<U>, L: LossFunction<U>>(
&mut self,
input: Self::LossInput,
stack: Self::OutStack,
optimizer: &mut OP,
lossf: &L
) -> Result<(), TrainingError>

fn backward_all<OP: Optimizer<U>, L: LossFunction<U>>(
&mut self,
input: Self::LossInput,
stack: Self::OutStack,
optimizer: &mut OP,
lossf: &L
) -> Result<(), TrainingError>

fn batch_backward<OP: Optimizer<U>, L: LossFunction<U>>(
&mut self,
input: Self::BatchLossInput,
stack: Self::BatchOutStack,
optimizer: &mut OP,
lossf: &L
) -> Result<(), TrainingError>

fn batch_backward<OP: Optimizer<U>, L: LossFunction<U>>(
&mut self,
input: Self::BatchLossInput,
stack: Self::BatchOutStack,
optimizer: &mut OP,
lossf: &L
) -> Result<(), TrainingError>

fn batch_forward(
&self,
input: Self::BatchInput
) -> Result<Self::BatchOutput, TrainingError>

fn batch_loss<L: LossFunction<U>>(
&self,
loss: Self::BatchLossInput,
_lossf: &L,
stack: Self::BatchOutStack
) -> Result<(Self::BatchOutStack, Self::BatchLossInput), TrainingError>

fn batch_loss<L: LossFunction<U>>(
&self,
loss: Self::BatchLossInput,
_lossf: &L,
stack: Self::BatchOutStack
) -> Result<(Self::BatchOutStack, Self::BatchLossInput), TrainingError>

fn batch_pre_train(
&self,
input: Self::BatchInput
) -> Result<Self::BatchOutStack, TrainingError>

fn loss<L: LossFunction<U>>(
&mut self,
loss: Self::LossInput,
_lossf: &L,
stack: Self::OutStack
) -> Result<(Self::OutStack, Self::LossInput), TrainingError>

fn loss<L: LossFunction<U>>(
&mut self,
loss: Self::LossInput,
_lossf: &L,
stack: Self::OutStack
) -> Result<(Self::OutStack, Self::LossInput), TrainingError>

fn load(
&mut self,
persistence: &mut TextFilePersistence<U>
) -> Result<(), ConfigReadError>

fn save(
&mut self,
persistence: &mut TextFilePersistence<U>
) -> Result<(), PersistenceError>

fn load(
&mut self,
persistence: &mut TextFilePersistence<U>
) -> Result<(), ConfigReadError>

fn save(
&mut self,
persistence: &mut TextFilePersistence<U>
) -> Result<(), PersistenceError>

impl<U, C, P, D, I, const NI: usize, const NO: usize> RefUnwindSafe for LinearLayer<U, C, P, D, I, NI, NO>where
C: RefUnwindSafe,
D: RefUnwindSafe,
P: RefUnwindSafe,
U: RefUnwindSafe,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Send for LinearLayer<U, C, P, D, I, NI, NO>where
C: Send,
D: Send,
P: Send,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Sync for LinearLayer<U, C, P, D, I, NI, NO>where
C: Sync,
D: Sync,
P: Sync,

impl<U, C, P, D, I, const NI: usize, const NO: usize> Unpin for LinearLayer<U, C, P, D, I, NI, NO>where
C: Unpin,
D: Unpin,
P: Unpin,

impl<U, C, P, D, I, const NI: usize, const NO: usize> UnwindSafe for LinearLayer<U, C, P, D, I, NI, NO>where
C: UnwindSafe,
D: UnwindSafe,
P: UnwindSafe,
U: UnwindSafe,

impl<T> AddLayer for Twhere
T: ForwardAll,

fn add_layer<C, F>(self, f: F) -> Cwhere
C: ForwardAll,
F: FnOnce(T) -> C,

impl<T, U> AddLayerTrain<U> for Twhere
T: PreTrain<U>,
U: UnitValue<U>,

fn add_layer_train<C, F>(self, f: F) -> Cwhere
C: Train<U>,
F: FnOnce(T) -> C,

impl<T> Any for Twhere
T: 'static + ?Sized,

impl<T> Borrow<T> for Twhere
T: ?Sized,

impl<T> BorrowMut<T> for Twhere
T: ?Sized,

impl<T, U> ForwardDiff<U> for Twhere
T: PreTrain<U>,
U: UnitValue<U>,

fn forward_diff(
&self,
input: <T as ForwardAll>::Input
) -> Result<<T as PreTrain<U>>::OutStack, EvaluateError>

impl<T, U> Into<U> for Twhere
U: From<T>,

impl<T> TryAddLayer for Twhere
T: ForwardAll,

fn try_add_layer<C, F>(self, f: F) -> Result<C, DeviceError>where
C: ForwardAll,
F: FnOnce(T) -> Result<C, DeviceError>,

impl<T, U> TryAddLayerTrain<U> for Twhere
T: PreTrain<U>,
U: UnitValue<U>,

fn try_add_layer_train<C, F>(self, f: F) -> Result<C, DeviceError>where
C: Train<U>,
F: FnOnce(T) -> Result<C, DeviceError>,

impl<T, U> TryFrom<U> for Twhere
U: Into<T>,

impl<T, U> TryInto<U> for Twhere
U: TryFrom<T>,