ImageModelDistributionSettings interface
Distribution expressions to sweep over values of model settings.
Some examples are:
ModelName = "choice('seresnext', 'resnest50')";
LearningRate = "uniform(0.001, 0.01)";
LayersToFreeze = "choice(0, 2)";
All distributions can be specified as distribution_name(min, max) or choice(val1, val2, ..., valn)
where distribution name can be: uniform, quniform, loguniform, etc
For more details on how to compose distribution expressions please check the documentation:
https://docs.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters
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.
Properties
| ams |
Enable AMSGrad when optimizer is 'adam' or 'adamw'. |
| augmentations | Settings for using Augmentations. |
| beta1 | Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
| beta2 | Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1]. |
| distributed | Whether to use distributer training. |
| early |
Enable early stopping logic during training. |
| early |
Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer. |
| early |
Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer. |
| enable |
Enable normalization when exporting ONNX model. |
| evaluation |
Frequency to evaluate validation dataset to get metric scores. Must be a positive integer. |
| gradient |
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. |
| layers |
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. |
| learning |
Initial learning rate. Must be a float in the range [0, 1]. |
| learning |
Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'. |
| model |
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. |
| momentum | Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1]. |
| nesterov | Enable nesterov when optimizer is 'sgd'. |
| number |
Number of training epochs. Must be a positive integer. |
| number |
Number of data loader workers. Must be a non-negative integer. |
| optimizer | Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'. |
| random |
Random seed to be used when using deterministic training. |
| step |
Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1]. |
| step |
Value of step size when learning rate scheduler is 'step'. Must be a positive integer. |
| training |
Training batch size. Must be a positive integer. |
| validation |
Validation batch size. Must be a positive integer. |
| warmup |
Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1]. |
| warmup |
Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer. |
| weight |
Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1]. |
Property Details
amsGradient
Enable AMSGrad when optimizer is 'adam' or 'adamw'.
amsGradient?: stringProperty Value
string
augmentations
Settings for using Augmentations.
augmentations?: stringProperty Value
string
beta1
Value of 'beta1' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
beta1?: stringProperty Value
string
beta2
Value of 'beta2' when optimizer is 'adam' or 'adamw'. Must be a float in the range [0, 1].
beta2?: stringProperty Value
string
distributed
Whether to use distributer training.
distributed?: stringProperty Value
string
earlyStopping
Enable early stopping logic during training.
earlyStopping?: stringProperty Value
string
earlyStoppingDelay
Minimum number of epochs or validation evaluations to wait before primary metric improvement is tracked for early stopping. Must be a positive integer.
earlyStoppingDelay?: stringProperty Value
string
earlyStoppingPatience
Minimum number of epochs or validation evaluations with no primary metric improvement before the run is stopped. Must be a positive integer.
earlyStoppingPatience?: stringProperty Value
string
enableOnnxNormalization
Enable normalization when exporting ONNX model.
enableOnnxNormalization?: stringProperty Value
string
evaluationFrequency
Frequency to evaluate validation dataset to get metric scores. Must be a positive integer.
evaluationFrequency?: stringProperty Value
string
gradientAccumulationStep
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.
gradientAccumulationStep?: stringProperty Value
string
layersToFreeze
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.
layersToFreeze?: stringProperty Value
string
learningRate
Initial learning rate. Must be a float in the range [0, 1].
learningRate?: stringProperty Value
string
learningRateScheduler
Type of learning rate scheduler. Must be 'warmup_cosine' or 'step'.
learningRateScheduler?: stringProperty Value
string
modelName
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.
modelName?: stringProperty Value
string
momentum
Value of momentum when optimizer is 'sgd'. Must be a float in the range [0, 1].
momentum?: stringProperty Value
string
nesterov
Enable nesterov when optimizer is 'sgd'.
nesterov?: stringProperty Value
string
numberOfEpochs
Number of training epochs. Must be a positive integer.
numberOfEpochs?: stringProperty Value
string
numberOfWorkers
Number of data loader workers. Must be a non-negative integer.
numberOfWorkers?: stringProperty Value
string
optimizer
Type of optimizer. Must be either 'sgd', 'adam', or 'adamw'.
optimizer?: stringProperty Value
string
randomSeed
Random seed to be used when using deterministic training.
randomSeed?: stringProperty Value
string
stepLRGamma
Value of gamma when learning rate scheduler is 'step'. Must be a float in the range [0, 1].
stepLRGamma?: stringProperty Value
string
stepLRStepSize
Value of step size when learning rate scheduler is 'step'. Must be a positive integer.
stepLRStepSize?: stringProperty Value
string
trainingBatchSize
Training batch size. Must be a positive integer.
trainingBatchSize?: stringProperty Value
string
validationBatchSize
Validation batch size. Must be a positive integer.
validationBatchSize?: stringProperty Value
string
warmupCosineLRCycles
Value of cosine cycle when learning rate scheduler is 'warmup_cosine'. Must be a float in the range [0, 1].
warmupCosineLRCycles?: stringProperty Value
string
warmupCosineLRWarmupEpochs
Value of warmup epochs when learning rate scheduler is 'warmup_cosine'. Must be a positive integer.
warmupCosineLRWarmupEpochs?: stringProperty Value
string
weightDecay
Value of weight decay when optimizer is 'sgd', 'adam', or 'adamw'. Must be a float in the range[0, 1].
weightDecay?: stringProperty Value
string