skainet-model-yolo/sk.ainet.lang.model.dnn.yolo/YoloConfig

YoloConfig

data class YoloConfig(val numClasses: Int = 80, val inputSize: Int = 640, val confThreshold: Float = 0.25f, val iouThreshold: Float = 0.45f, val maxDetections: Int = 300, val classNames: List<String> = emptyList(), val regMax: Int = 16, val baseChannels: Int = 32, val depthMultiple: Float = 1.0f)(source)

YOLO configuration used for decoding and thresholds.

Constructors

YoloConfig

constructor(numClasses: Int = 80, inputSize: Int = 640, confThreshold: Float = 0.25f, iouThreshold: Float = 0.45f, maxDetections: Int = 300, classNames: List<String> = emptyList(), regMax: Int = 16, baseChannels: Int = 32, depthMultiple: Float = 1.0f)

Properties

baseChannels

val baseChannels: Int

Channel multiplier used to size the backbone/neck to match ONNX weights. Example: YOLOv8n exported from Ultralytics uses baseChannels=16 (width_mult=0.25), while larger configs use higher numbers.

classNames

val classNames: List<String>

confThreshold

val confThreshold: Float

depthMultiple

val depthMultiple: Float

Depth multiplier controlling how many bottlenecks each C2f block uses. Example: Ultralytics YOLOv8n uses ~0.33 which turns base counts 3,6,6,3 into 1,2,2,1.

inputSize

val inputSize: Int

iouThreshold

val iouThreshold: Float

maxDetections

val maxDetections: Int

numClasses

val numClasses: Int

regMax

val regMax: Int