org.apache.ignite.tensorflow.cluster.util

Class TensorFlowChiefRunner

java.lang.Object

org.apache.ignite.tensorflow.core.util.AsyncNativeProcessRunner

org.apache.ignite.tensorflow.cluster.util.TensorFlowChiefRunner

public class TensorFlowChiefRunner
extends AsyncNativeProcessRunner

Utils class that helps to start and stop chief process.

Constructor Summary

Constructors

Constructor and Description
TensorFlowChiefRunner(Ignite ignite, ExecutorService executor, TensorFlowClusterSpec spec, Consumer<String> out, Consumer<String> err) Constructs a new instance of TensorFlow chief runner.

Method Summary

Modifier and Type	Method and Description
void	doAfter() Method that should be called after starting the process.
NativeProcessRunner	doBefore() Method that should be called before starting the process.

Methods inherited from class org.apache.ignite.tensorflow.core.util.AsyncNativeProcessRunner

getException, isCompleted, start, stop

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

TensorFlowChiefRunner

public TensorFlowChiefRunner(Ignite ignite,
                             ExecutorService executor,
                             TensorFlowClusterSpec spec,
                             Consumer<String> out,
                             Consumer<String> err)

Constructs a new instance of TensorFlow chief runner.

Parameters:

ignite - Ignite instance.

executor - Executor to be used in AsyncNativeProcessRunner.

spec - TensorFlow cluster specification.

out - Output stream data consumer.

err - Error stream data consumer.

Method Detail

doBefore

public NativeProcessRunner doBefore()

Method that should be called before starting the process.

Specified by:

doBefore in class AsyncNativeProcessRunner

Returns:

Prepared native process runner.

doAfter

public void doAfter()

Method that should be called after starting the process.

Specified by:

doAfter in class AsyncNativeProcessRunner