TrainingJobAlgorithmSpecification constructor

const TrainingJobAlgorithmSpecification({
  1. Input<String?>? algorithmName,
  2. Input<List<String>?>? containerArguments,
  3. Input<List<String>?>? containerEntrypoints,
  4. Input<bool?>? enableSagemakerMetricsTimeSeries,
  5. Input<List<TrainingJobAlgorithmSpecificationMetricDefinition>?>? metricDefinitions,
  6. Input<String?>? trainingImage,
  7. Input<TrainingJobAlgorithmSpecificationTrainingImageConfig?>? trainingImageConfig,
  8. Input<String?>? trainingInputMode,
})

Creates a new TrainingJobAlgorithmSpecification. algorithmName Name or ARN of the algorithm resource to use for the training job. containerArguments List of arguments for the container entrypoint. Maximum of 100 entries. containerEntrypoints List of entrypoint commands for the container. Maximum of 100 entries. enableSagemakerMetricsTimeSeries Whether to enable SageMaker AI metrics time series collection. metricDefinitions List of metric definitions for the training job. Maximum of 40. Use this to extract custom metrics from your own training container logs. SageMaker can still publish built-in metrics for built-in algorithms and supported prebuilt images when this block is omitted. See metricDefinitions below. trainingImage Registry path of the Docker image that contains the training algorithm. trainingImageConfig Training image configuration. See trainingImageConfig below. trainingInputMode Input mode for the training data. Valid values: File, Pipe, FastFile.

Implementation

const TrainingJobAlgorithmSpecification({
  this.algorithmName,
  this.containerArguments,
  this.containerEntrypoints,
  this.enableSagemakerMetricsTimeSeries,
  this.metricDefinitions,
  this.trainingImage,
  this.trainingImageConfig,
  this.trainingInputMode,
});