Class | Description |
---|---|
ExecuteWorkerFlatMap<R extends TrainingResult> |
A FlatMapFunction for executing training on DataSets.
|
ExecuteWorkerMultiDataSetFlatMap<R extends TrainingResult> |
A FlatMapFunction for executing training on MultiDataSets.
|
ExecuteWorkerPathFlatMap<R extends TrainingResult> |
A FlatMapFunction for executing training on serialized DataSet objects, that can be loaded from a path (local or HDFS)
that is specified as a String
Used in both SparkDl4jMultiLayer and SparkComputationGraph implementations
|
ExecuteWorkerPathMDSFlatMap<R extends TrainingResult> |
A FlatMapFunction for executing training on serialized DataSet objects, that can be loaded from a path (local or HDFS)
that is specified as a String
Used in both SparkDl4jMultiLayer and SparkComputationGraph implementations
|
ExecuteWorkerPDSFlatMap<R extends TrainingResult> |
A FlatMapFunction for executing training on serialized DataSet objects, that can be loaded using a PortableDataStream
Used in both SparkDl4jMultiLayer and SparkComputationGraph implementations
|
ExecuteWorkerPDSMDSFlatMap<R extends TrainingResult> |
A FlatMapFunction for executing training on serialized MultiDataSet objects, that can be loaded using a PortableDataStream
Used for SparkComputationGraph implementations only
|
NetBroadcastTuple |
A simple class for storing configurations, parameters and updaters in one class (so they can be broadcast together)
|