Options
All
  • Public
  • Public/Protected
  • All
Menu

Namespace TFewTrainingConfig

The TFEW training method hyperparameters.

Properties

Methods

Properties

Optional earlyStoppingPatience

earlyStoppingPatience: undefined | number

Stop training if the loss metric does not improve beyond 'early_stopping_threshold' for this many times of evaluation. Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.

Optional earlyStoppingThreshold

earlyStoppingThreshold: undefined | number

How much the loss must improve to prevent early stopping. Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.

Optional learningRate

learningRate: undefined | number

The initial learning rate to be used during training Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.

Optional logModelMetricsIntervalInSteps

logModelMetricsIntervalInSteps: undefined | number

Determines how frequently to log model metrics.

Every step is logged for the first 20 steps and then follows this parameter for log frequency. Set to 0 to disable logging the model metrics. Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.

Optional totalTrainingEpochs

totalTrainingEpochs: undefined | number

The maximum number of training epochs to run for. Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.

Optional trainingBatchSize

trainingBatchSize: undefined | number

The batch size used during training. Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.

Const trainingConfigType

trainingConfigType: string = "TFEW_TRAINING_CONFIG"

Methods

getDeserializedJsonObj

getJsonObj