ImageModelSettingsObjectDetectionArgs

data class ImageModelSettingsObjectDetectionArgs(val advancedSettings: Output<String>? = null, val amsGradient: Output<Boolean>? = null, val augmentations: Output<String>? = null, val beta1: Output<Double>? = null, val beta2: Output<Double>? = null, val boxDetectionsPerImage: Output<Int>? = null, val boxScoreThreshold: Output<Double>? = null, val checkpointFrequency: Output<Int>? = null, val checkpointModel: Output<MLFlowModelJobInputArgs>? = null, val checkpointRunId: Output<String>? = null, val distributed: Output<Boolean>? = null, val earlyStopping: Output<Boolean>? = null, val earlyStoppingDelay: Output<Int>? = null, val earlyStoppingPatience: Output<Int>? = null, val enableOnnxNormalization: Output<Boolean>? = null, val evaluationFrequency: Output<Int>? = null, val gradientAccumulationStep: Output<Int>? = null, val imageSize: Output<Int>? = null, val layersToFreeze: Output<Int>? = null, val learningRate: Output<Double>? = null, val learningRateScheduler: Output<Either<String, LearningRateScheduler>>? = null, val maxSize: Output<Int>? = null, val minSize: Output<Int>? = null, val modelName: Output<String>? = null, val modelSize: Output<Either<String, ModelSize>>? = null, val momentum: Output<Double>? = null, val multiScale: Output<Boolean>? = null, val nesterov: Output<Boolean>? = null, val nmsIouThreshold: Output<Double>? = null, val numberOfEpochs: Output<Int>? = null, val numberOfWorkers: Output<Int>? = null, val optimizer: Output<Either<String, StochasticOptimizer>>? = null, val randomSeed: Output<Int>? = null, val stepLRGamma: Output<Double>? = null, val stepLRStepSize: Output<Int>? = null, val tileGridSize: Output<String>? = null, val tileOverlapRatio: Output<Double>? = null, val tilePredictionsNmsThreshold: Output<Double>? = null, val trainingBatchSize: Output<Int>? = null, val validationBatchSize: Output<Int>? = null, val validationIouThreshold: Output<Double>? = null, val validationMetricType: Output<Either<String, ValidationMetricType>>? = null, val warmupCosineLRCycles: Output<Double>? = null, val warmupCosineLRWarmupEpochs: Output<Int>? = null, val weightDecay: Output<Double>? = null) : ConvertibleToJava<ImageModelSettingsObjectDetectionArgs>

Settings used for training the model. For more information on the available settings please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.

Constructors

Link copied to clipboard
constructor(advancedSettings: Output<String>? = null, amsGradient: Output<Boolean>? = null, augmentations: Output<String>? = null, beta1: Output<Double>? = null, beta2: Output<Double>? = null, boxDetectionsPerImage: Output<Int>? = null, boxScoreThreshold: Output<Double>? = null, checkpointFrequency: Output<Int>? = null, checkpointModel: Output<MLFlowModelJobInputArgs>? = null, checkpointRunId: Output<String>? = null, distributed: Output<Boolean>? = null, earlyStopping: Output<Boolean>? = null, earlyStoppingDelay: Output<Int>? = null, earlyStoppingPatience: Output<Int>? = null, enableOnnxNormalization: Output<Boolean>? = null, evaluationFrequency: Output<Int>? = null, gradientAccumulationStep: Output<Int>? = null, imageSize: Output<Int>? = null, layersToFreeze: Output<Int>? = null, learningRate: Output<Double>? = null, learningRateScheduler: Output<Either<String, LearningRateScheduler>>? = null, maxSize: Output<Int>? = null, minSize: Output<Int>? = null, modelName: Output<String>? = null, modelSize: Output<Either<String, ModelSize>>? = null, momentum: Output<Double>? = null, multiScale: Output<Boolean>? = null, nesterov: Output<Boolean>? = null, nmsIouThreshold: Output<Double>? = null, numberOfEpochs: Output<Int>? = null, numberOfWorkers: Output<Int>? = null, optimizer: Output<Either<String, StochasticOptimizer>>? = null, randomSeed: Output<Int>? = null, stepLRGamma: Output<Double>? = null, stepLRStepSize: Output<Int>? = null, tileGridSize: Output<String>? = null, tileOverlapRatio: Output<Double>? = null, tilePredictionsNmsThreshold: Output<Double>? = null, trainingBatchSize: Output<Int>? = null, validationBatchSize: Output<Int>? = null, validationIouThreshold: Output<Double>? = null, validationMetricType: Output<Either<String, ValidationMetricType>>? = null, warmupCosineLRCycles: Output<Double>? = null, warmupCosineLRWarmupEpochs: Output<Int>? = null, weightDecay: Output<Double>? = null)

Properties

Link copied to clipboard
val advancedSettings: Output<String>? = null

Settings for advanced scenarios.

Link copied to clipboard
val amsGradient: Output<Boolean>? = null

Enable AMSGrad when optimizer is 'adam' or 'adamw'.

Link copied to clipboard
val augmentations: Output<String>? = null

Settings for using Augmentations.

Link copied to clipboard
val beta1: Output<Double>? = null

Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range 0, 1.

Link copied to clipboard
val beta2: Output<Double>? = null

Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range 0, 1.

Link copied to clipboard
val boxDetectionsPerImage: Output<Int>? = null

Maximum number of detections per image, for all classes. Must be a positive integer. Note: This settings is not supported for the 'yolov5' algorithm.

Link copied to clipboard
val boxScoreThreshold: Output<Double>? = null

During inference, only return proposals with a classification score greater than BoxScoreThreshold. Must be a float in the range0, 1.

Link copied to clipboard
val checkpointFrequency: Output<Int>? = null

Frequency to store model checkpoints. Must be a positive integer.

Link copied to clipboard

The pretrained checkpoint model for incremental training.

Link copied to clipboard
val checkpointRunId: Output<String>? = null

The id of a previous run that has a pretrained checkpoint for incremental training.

Link copied to clipboard
val distributed: Output<Boolean>? = null

Whether to use distributed training.

Link copied to clipboard
val earlyStopping: Output<Boolean>? = null

Enable early stopping logic during training.

Link copied to clipboard
val earlyStoppingDelay: Output<Int>? = null

Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer.

Link copied to clipboard
val earlyStoppingPatience: Output<Int>? = null

Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer.

Link copied to clipboard
val enableOnnxNormalization: Output<Boolean>? = null

Enable normalization when exporting ONNX model.

Link copied to clipboard
val evaluationFrequency: Output<Int>? = null

Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.

Link copied to clipboard
val gradientAccumulationStep: Output<Int>? = null

Gradient accumulation means running a configured number of "GradAccumulationStep" steps without updating the model weights while accumulating the gradients of those steps, and then using the accumulated gradients to compute the weight updates. Must be a positive integer.

Link copied to clipboard
val imageSize: Output<Int>? = null

Image size for train and validation. Must be a positive integer. Note: The training run may get into CUDA OOM if the size is too big. Note: This settings is only supported for the 'yolov5' algorithm.

Link copied to clipboard
val layersToFreeze: Output<Int>? = null

Number of layers to freeze for the model. Must be a positive integer. For instance, passing 2 as value for 'seresnext' means freezing layer0 and layer1. For a full list of models supported and details on layer freeze, please see: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.

Link copied to clipboard
val learningRate: Output<Double>? = null

Initial learning rate. Must be a float in the range 0, 1.

Link copied to clipboard
val learningRateScheduler: Output<Either<String, LearningRateScheduler>>? = null

Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.

Link copied to clipboard
val maxSize: Output<Int>? = null

Maximum size of the image to be rescaled before feeding it to the backbone. Must be a positive integer. Note: training run may get into CUDA OOM if the size is too big. Note: This settings is not supported for the 'yolov5' algorithm.

Link copied to clipboard
val minSize: Output<Int>? = null

Minimum size of the image to be rescaled before feeding it to the backbone. Must be a positive integer. Note: training run may get into CUDA OOM if the size is too big. Note: This settings is not supported for the 'yolov5' algorithm.

Link copied to clipboard
val modelName: Output<String>? = null

Name of the model to use for training. For more information on the available models please visit the official documentation: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-auto-train-image-models.

Link copied to clipboard
val modelSize: Output<Either<String, ModelSize>>? = null

Model size. Must be 'small', 'medium', 'large', or 'xlarge'. Note: training run may get into CUDA OOM if the model size is too big. Note: This settings is only supported for the 'yolov5' algorithm.

Link copied to clipboard
val momentum: Output<Double>? = null

Value of momentum when optimizer is 'sgd'. Must be a float in the range 0, 1.

Link copied to clipboard
val multiScale: Output<Boolean>? = null

Enable multi-scale image by varying image size by +/- 50%. Note: training run may get into CUDA OOM if no sufficient GPU memory. Note: This settings is only supported for the 'yolov5' algorithm.

Link copied to clipboard
val nesterov: Output<Boolean>? = null

Enable nesterov when optimizer is 'sgd'.

Link copied to clipboard
val nmsIouThreshold: Output<Double>? = null

IOU threshold used during inference in NMS post processing. Must be a float in the range 0, 1.

Link copied to clipboard
val numberOfEpochs: Output<Int>? = null

Number of training epochs. Must be a positive integer.

Link copied to clipboard
val numberOfWorkers: Output<Int>? = null

Number of data loader workers. Must be a non-negative integer.

Link copied to clipboard
val optimizer: Output<Either<String, StochasticOptimizer>>? = null

Type of optimizer.

Link copied to clipboard
val randomSeed: Output<Int>? = null

Random seed to be used when using deterministic training.

Link copied to clipboard
val stepLRGamma: Output<Double>? = null

Value of gamma when learning rate scheduler is 'step'. Must be a float in the range 0, 1.

Link copied to clipboard
val stepLRStepSize: Output<Int>? = null

Value of step size when learning rate scheduler is 'step'. Must be a positive integer.

Link copied to clipboard
val tileGridSize: Output<String>? = null

The grid size to use for tiling each image. Note: TileGridSize must not be None to enable small object detection logic. A string containing two integers in mxn format. Note: This settings is not supported for the 'yolov5' algorithm.

Link copied to clipboard
val tileOverlapRatio: Output<Double>? = null

Overlap ratio between adjacent tiles in each dimension. Must be float in the range [0, 1). Note: This settings is not supported for the 'yolov5' algorithm.

Link copied to clipboard
val tilePredictionsNmsThreshold: Output<Double>? = null

The IOU threshold to use to perform NMS while merging predictions from tiles and image. Used in validation/ inference. Must be float in the range 0, 1. Note: This settings is not supported for the 'yolov5' algorithm.

Link copied to clipboard
val trainingBatchSize: Output<Int>? = null

Training batch size. Must be a positive integer.

Link copied to clipboard
val validationBatchSize: Output<Int>? = null

Validation batch size. Must be a positive integer.

Link copied to clipboard
val validationIouThreshold: Output<Double>? = null

IOU threshold to use when computing validation metric. Must be float in the range 0, 1.

Link copied to clipboard
val validationMetricType: Output<Either<String, ValidationMetricType>>? = null

Metric computation method to use for validation metrics.

Link copied to clipboard
val warmupCosineLRCycles: Output<Double>? = null

Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range 0, 1.

Link copied to clipboard
val warmupCosineLRWarmupEpochs: Output<Int>? = null

Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.

Link copied to clipboard
val weightDecay: Output<Double>? = null

Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range0, 1.

Functions

Link copied to clipboard
open override fun toJava(): ImageModelSettingsObjectDetectionArgs