BiasFreeLinearLayer<T>
Layers ยท AiDotNet.NeuralNetworks.Layers
Trainable linear projection y = xW with no bias term.
Inheritance: LayerBase<T> โ BiasFreeLinearLayer<T>
Implements: IShapeContract
For Beginners
This layer changes the feature width using only a weight matrix. Omitting the bias is useful when the surrounding architecture already supplies an offset or when the paper explicitly requires a pure projection.
Constructors
| Constructor | Summary |
|---|---|
BiasFreeLinearLayer(int inputSize, int outputSize) |
Properties
| Property | Summary |
|---|---|
bool SupportsTraining { get; } |
|
int InputSize { get; } |
|
int OutputSize { get; } |
Methods
| Method | Summary |
|---|---|
Dictionary<string, string> GetParameterRoles() |
Returns parameter roles for per-role optimizer configuration (e.g., weight decay exemption for biases). |
IReadOnlyList<OutputAxisContract> OutputAxesFor(int inputRank) |
|
IReadOnlyList<Tensor<T>> GetTrainableParameters() |
Returns all trainable parameter tensors marked with [TrainableParameter]. |
void ResetState() |
|
void SetTrainableParameters(IReadOnlyList<Tensor<T>> parameters) |
Replaces trainable parameter tensors (e.g., with ParameterBuffer views). |
void UpdateParameters(T learningRate) |
|
void ZeroGrad() |
Clears all gradient fields discovered by convention ({paramName}Gradient). |