HiFiGANResBlockLayer<T>
Layers · AiDotNet.NeuralNetworks.Layers
HiFi-GAN Multi-Receptive Field (MRF) fusion module (Kong et al. 2020, §2.2) for
1-D waveform/feature data [B, C, T]. After each upsampling stage the
generator runs the input through several residual blocks with different kernel
sizes and dilation patterns IN PARALLEL and returns their averaged sum, so the
network observes patterns over diverse receptive fields simultaneously:
Inheritance: LayerBase<T> → HiFiGANResBlockLayer<T>
Implements: IShapeContract
How It Works
The official jik876/hifi-gan v1 config uses
resblock_kernel_sizes=[3,7,11] and
resblock_dilation_sizes=[[1,3,5],[1,3,5],[1,3,5]] — the defaults here. The
parallel-branch SUM is the defining MRF behaviour; a single sequential dilated-conv
chain is NOT MRF.
Built from inner Conv1DLayer instances (two per dilation per
kernel — the leaky-pre-activated dilated conv and the dilation-1 projection), so
the tape handles backward and the fused conv kernels are reused. “Same” padding
keeps T constant across the block (required for the per-branch residual adds and
the cross-branch sum). Reconstructable from (channels, kernelSizes, dilations).
Constructors
| Constructor | Summary |
|---|---|
HiFiGANResBlockLayer(int channels, int[] kernelSizes, int[] dilations) |
Constructs a HiFi-GAN MRF block over the given kernel sizes / dilations. |
Properties
| Property | Summary |
|---|---|
bool SupportsTraining { get; } |
Methods
| Method | Summary |
|---|---|
IReadOnlyList<ILayer<T>> GetSubLayers() |
|
IReadOnlyList<OutputAxisContract> OutputAxesFor(int inputRank) |
Derived from this type’s [TensorLayout] declarations. |
void ResetState() |
|
void SetTrainingMode(bool isTraining) |
|
void UpdateParameters(T learningRate) |