GroundingVLMOptions

Options & Configuration · AiDotNet.VisionLanguage.Grounding

Base configuration options for visual grounding models.

Inheritance: ModelOptions → GenerativeVLMOptions → GroundingVLMOptions

Known derived types: DINOXOptions, FerretOptions, FerretV2Options, GromaOptions, GroundingDINO15Options, GroundingDINOOptions, OWLViTOptions, OWLv2Options, ShikraOptions

For Beginners

These options configure the Grounding model. Default values follow the original paper settings.

Constructors

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

Properties

Property Summary
double ConfidenceThreshold { get; set; } Gets or sets the confidence threshold for detection filtering.
double NmsThreshold { get; set; } Gets or sets the IoU threshold for non-maximum suppression.
int BoxDimension { get; set; } Gets or sets the number of output coordinates per box (typically 4 for [x1,y1,x2,y2]).
int MaxDetections { get; set; } Gets or sets the maximum number of detections per image.