Package ai.djl.modality.cv
Class MultiBoxTarget.Builder
java.lang.Object
ai.djl.modality.cv.MultiBoxTarget.Builder
- Enclosing class:
- MultiBoxTarget
The Builder to construct a
MultiBoxTarget object.-
Method Summary
Modifier and TypeMethodDescriptionbuild()Builds aMultiBoxTargetblock.optIgnoreLabel(float ignoreLabel) Sets the label for ignored anchors.optIouThreshold(float iouThreshold) Sets the anchor-GroundTruth overlap threshold to be regarded as a positive match.optMinNegativeSamples(int minNegativeSamples) Sets the minimum number of negative samples.optNegativeMinigRatio(float negativeMinigRatio) Sets the max negative to positive samples ratio.optNegativeMiningThreshold(float negativeMiningThreshold) Sets the threshold used for negative mining.
-
Method Details
-
optMinNegativeSamples
Sets the minimum number of negative samples.- Parameters:
minNegativeSamples- the minimum number of negative samples- Returns:
- this
Builder
-
optIouThreshold
Sets the anchor-GroundTruth overlap threshold to be regarded as a positive match.- Parameters:
iouThreshold- the anchor-GroundTruth overlap threshold to be regarded as a positive match- Returns:
- this
Builder
-
optIgnoreLabel
Sets the label for ignored anchors. Defaults to -1.- Parameters:
ignoreLabel- the label for ignored anchors- Returns:
- this
Builder
-
optNegativeMiningThreshold
Sets the threshold used for negative mining.- Parameters:
negativeMiningThreshold- the threshold used for negative mining- Returns:
- this
Builder
-
optNegativeMinigRatio
Sets the max negative to positive samples ratio. Use -1 to disable mining. Defaults to -1.- Parameters:
negativeMinigRatio- the max negative to positive samples ratio- Returns:
- this
Builder
-
build
Builds aMultiBoxTargetblock.- Returns:
- the
MultiBoxTargetblock
-