OneFormerOptions

Options & Configuration · AiDotNet.ComputerVision.Segmentation.Foundation

Configuration options for the OneFormer universal segmentation model.

Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → PanopticSegmentationOptions → OneFormerOptions

For Beginners

OneFormer options inherit from NeuralNetworkOptions. OneFormer is trained once on panoptic data and can perform semantic, instance, or panoptic segmentation by simply providing a text prompt describing which task to perform. This “one model, all tasks” approach simplifies deployment.

Constructors

Constructor Summary
OneFormerOptions() Initializes a new instance with default values.
OneFormerOptions(OneFormerOptions other) Initializes a new instance by copying from another instance.

Properties

Property Summary
OneFormerModelSize ModelSize { get; set; } Gets or sets the backbone size variant.
double LearningRate { get; set; } Gets or sets the AdamW learning rate.
double MaxGradientNorm { get; set; } Gets or sets the global gradient-norm clipping threshold.
double WeightDecay { get; set; } Gets or sets the decoupled AdamW weight decay.
int MlpRatio { get; set; } Gets or sets the Swin MLP expansion ratio.
int PatchSize { get; set; } Gets or sets the Swin patch size.
int WindowSize { get; set; } Gets or sets the Swin attention window size.
int? DecoderDimension { get; set; } Gets or sets the decoder width.
int[] AttentionHeads { get; set; } Gets or sets the four Swin attention-head counts.
int[] ChannelDimensions { get; set; } Gets or sets the four Swin stage dimensions.
int[] StageDepths { get; set; } Gets or sets the four Swin stage depths.

Methods

Method Summary
void Validate() Throws if a value this model requires has been left unset or is not positive.