Package org.apache.ignite.tensorflow.cluster

TensorFlow integration API that allows to start and maintain TensorFlow cluster using infrastructure tools from package org.apache.ignite.tensorflow.core.

See: Description

Class Summary

Class	Description
TensorFlowCluster	TensorFlow Cluster metadata.
TensorFlowClusterGateway	TensorFlow cluster gateway that allows to subscribe on changes in cluster configuration.
TensorFlowClusterGatewayManager	TensorFlow cluster manager that allows to start, maintain and stop TensorFlow cluster using TensorFlowClusterManager and Ignite Service Grid.
TensorFlowClusterMaintainer	TensorFlow cluster service that maintains TensorFlow cluster.
TensorFlowClusterManager	TensorFlow cluster manager that allows to start, maintain and stop TensorFlow cluster.
TensorFlowJobArchive	TensorFlow job archive that keeps archived working directory and command to be executed.

Package org.apache.ignite.tensorflow.cluster Description

TensorFlow integration API that allows to start and maintain TensorFlow cluster using infrastructure tools from package org.apache.ignite.tensorflow.core. The most important components are:

• TensorFlowClusterManager that allows to start and stop TensorFlow cluster on top of Apache Ignite, but doesn't monitor it and doesn't maintain so that in case of failure the cluster won't be restarted.
• TensorFlowClusterGatewayManager that allows to start, maintain and stop TensorFlow cluster on top of Apache Ignite so that in case of failure the cluster will be restarted and recovered.
• TensorFlowClusterGateway that allows to subscribe on cluster configuration changes that might be done as result of rebalancing or node failures.