UPRNetOptions
Options & Configuration · AiDotNet.Video.Options
Configuration options for UPR-Net unified pyramid recurrent network.
Inheritance: ModelOptions → NeuralNetworkOptions → UPRNetOptions
For Beginners
UPR-Net combines motion estimation and frame creation into a single efficient network that processes images at multiple scales. At each scale, it repeatedly refines its predictions until they’re good enough, like iterating on a drawing.
How It Works
UPR-Net (Jin et al., 2023) recurrently reuses one feature pyramid, one partial-correlation motion estimator, and one synthesis network across image-pyramid levels. The recurrence is across levels; the released architecture has no ConvLSTM and no per-level copies of the motion or synthesis weights.
Constructors
| Constructor | Summary |
|---|---|
UPRNetOptions() |
Initializes a new instance with default values. |
UPRNetOptions(UPRNetOptions other) |
Initializes a new instance by copying from another instance. |
Properties
| Property | Summary |
|---|---|
OnnxModelOptions OnnxOptions { get; set; } |
Gets or sets the ONNX runtime options. |
VideoModelVariant Variant { get; set; } |
Gets or sets the model variant. |
double DropoutRate { get; set; } |
Gets or sets the dropout rate. |
double LearningRate { get; set; } |
Gets or sets the learning rate. |
int NumLevelsSkipped { get; set; } |
Gets or sets the number of finest pyramid levels whose motion estimation is skipped. |
int NumPyramidLevels { get; set; } |
Gets or sets the number of pyramid levels. |
string ModelPath { get; set; } |
Gets or sets the path to the ONNX model file. |
Methods
| Method | Summary |
|---|---|
void Validate() |
Validates paper-architecture options. |