Layer<T>

Layers · AiDotNet.Finance.Trading.Factors

One stacked encoder layer’s blocks. The paper stacks L of these; the reference config sets layers = 2.

Implements: IEquatable<Layer<T>>

Constructors

Constructor Summary
Layer(StockformerAttention<T> LowTemporal, StockformerAttention<T> LowSpatial, StockformerAttention<T> HighSpatial, ILayer<T> HighTemporalProjection) One stacked encoder layer’s blocks.

Properties

Property Summary
ILayer<T> HighTemporalProjection { get; set; } Projection after the high band’s causal convolution — its temporal operator is convolutional, per the reference, so it needs no Q/K/V triple.
StockformerAttention<T> HighSpatial { get; set; } Eq.
StockformerAttention<T> LowSpatial { get; set; } Eq.
StockformerAttention<T> LowTemporal { get; set; } Eq.

Methods

Method Summary
Layer<T> <Clone>$()
bool Equals(Layer<T> other)
bool Equals(object obj)
bool op_Equality(Layer<T> left, Layer<T> right)
bool op_Inequality(Layer<T> left, Layer<T> right)
int GetHashCode()
string ToString()
void Deconstruct(StockformerAttention<T> LowTemporal, StockformerAttention<T> LowSpatial, StockformerAttention<T> HighSpatial, ILayer<T> HighTemporalProjection)