ViTObjectDetector class

A complete Vision Transformer-based Object Detector.

This model combines a ViT backbone for feature extraction with a simple detection head for predicting a fixed number of bounding boxes and class labels.

Note: This is a highly simplified architecture for demonstration. Real-world object detectors like DETR, Faster R-CNN, YOLO are significantly more complex, involving multi-scale features, sophisticated heads, and specialized loss functions.

Inheritance

• Object
• Module
• ViTObjectDetector

Constructors

ViTObjectDetector({required int imageSize, required int patchSize, required int numChannels, required int embedSize, required int numLayers, required int numHeads, required int numClasses, required int numQueries})

Properties

backbone → ViTBackbone

final

detectionHead → ObjectDetectionHead

final

hashCode → int

The hash code for this object.

no setter, inherited

numQueries → int

final

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

Methods

forward(List<double> imageData) → Map<String, List<ValueVector>>

Forward pass for the object detector.

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

parameters() → List<Value>

override

toString() → String

A string representation of this object.

inherited

zeroGrad() → void

inherited

Operators

operator ==(Object other) → bool

The equality operator.

inherited