Class MultiViewUNet<T>

Namespace

AiDotNet.Diffusion.Models

Assembly

AiDotNet.dll

Multi-view aware U-Net for MVDream.

public class MultiViewUNet<T>

Type Parameters

T

Numeric type.

Inheritance

object

MultiViewUNet<T>

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Constructors

MultiViewUNet(int, int, int, int, int, int?)

Creates a new multi-view U-Net.

public MultiViewUNet(int inputChannels, int outputChannels, int baseChannels, int numViews, int contextDim, int? seed = null)

Parameters

inputChannels int

outputChannels int

baseChannels int

numViews int

contextDim int

seed int?

Properties

BaseUNet

Gets the base UNet for interface compatibility.

public UNetNoisePredictor<T> BaseUNet { get; }

Property Value

UNetNoisePredictor<T>

ParameterCount

Gets parameter count.

public int ParameterCount { get; }

Property Value

int

SupportsCFG

Gets whether classifier-free guidance is supported.

public bool SupportsCFG { get; }

Property Value

bool

Methods

GetParameters()

Gets parameters.

public Vector<T> GetParameters()

Returns

Vector<T>

PredictNoise(Tensor<T>, int, Tensor<T>?)

Predicts noise for a single view.

public Tensor<T> PredictNoise(Tensor<T> input, int timestep, Tensor<T>? conditioning = null)

Parameters

input Tensor<T>

timestep int

conditioning Tensor<T>

Returns

Tensor<T>

PredictNoiseMultiView(Tensor<T>, int, Tensor<T>[])

Predicts noise for multiple views with cross-view attention.

public Tensor<T> PredictNoiseMultiView(Tensor<T> multiViewInput, int timestep, Tensor<T>[] viewConditions)

Parameters

multiViewInput Tensor<T>

timestep int

viewConditions Tensor<T>[]

Returns

Tensor<T>

SetParameters(Vector<T>)

Sets parameters.

public void SetParameters(Vector<T> parameters)

Parameters

parameters Vector<T>