nz.co.jedsimson.lgp.core.evolution.training

Package nz.co.jedsimson.lgp.core.evolution.training

Types

DistributedTrainer

class DistributedTrainer<TProgram, TOutput : Output<TProgram>> : Trainer<TProgram, TOutput, ProgressUpdate<TProgram, TOutput>>

Trains the model for a given number of runs, in a parallel manner.

DistributedTrainingJob

class DistributedTrainingJob<TProgram, TOutput : Output<TProgram>> : TrainingJob<TProgram, TOutput, ProgressUpdate<TProgram, TOutput>>

Represents an asynchronous distributed training operation.

SequentialTrainer

class SequentialTrainer<TProgram, TOutput : Output<TProgram>> : Trainer<TProgram, TOutput, ProgressUpdate<TProgram, TOutput>>

Trains the model for a given number of runs, in a sequential manner.

SequentialTrainingJob

class SequentialTrainingJob<TProgram, TOutput : Output<TProgram>> : TrainingJob<TProgram, TOutput, ProgressUpdate<TProgram, TOutput>>

Represents an asynchronous sequential training operation.

Trainer

abstract class Trainer<TProgram, TOutput : Output<TProgram>, TMessage : TrainingUpdateMessage>

A service capable of training evolutionary models in a particular environment.

TrainingJob

abstract class TrainingJob<TProgram, TOutput : Output<TProgram>, TMessage : TrainingUpdateMessage>

Represents an asynchronous training operation.

TrainingMessages

object TrainingMessages

Contains the different messages that can be sent from a Trainer when training asynchronously.

TrainingResult

data class TrainingResult<TProgram, TOutput : Output<TProgram>>

Represents the result of training a model using a runner.

TrainingUpdateMessage

interface TrainingUpdateMessage

A message sent from a Trainer to a subscriber (by calling TrainingJob.subscribeToUpdates.