nil_sink¶
Module: missions.nodes.sink.nil_sink
¶
Store only meta data of dataset
This is useful to save disk space in scenarios, where only the output of any preceding node is desired, and where there is no need to save the whole data set again.
Inheritance diagram for pySPACE.missions.nodes.sink.nil_sink
:
Class Summary¶
NilSinkNode (\*\*kwargs) |
Store only meta information and perform training and testing |
OnlyTrainSinkNode (\*\*kwargs) |
Store only meta information and perform training but not testing |
Classes¶
NilSinkNode
¶
-
class
pySPACE.missions.nodes.sink.nil_sink.
NilSinkNode
(**kwargs)[source]¶ Bases:
pySPACE.missions.nodes.sink.time_series_sink.TimeSeriesSinkNode
Store only meta information and perform training and testing
The node inherits from TimeSeriesSinkNode, but instead of collecting the data, this node passes a DummyDataset.
Parameters
Exemplary Call
- node: Nil_Sink
Input: any
Output: DummyDataset
Author: David Feess (david.feess@dfki.de)
Created: 2010/03/30
POSSIBLE NODE NAMES: - NilSink
- NilSinkNode
- Nil_Sink
POSSIBLE INPUT TYPES: - TimeSeries
- FeatureVector
- PredictionVector
Class Components Summary
get_result_dataset
()Return the empty dummy collection input_types
process_current_split
()Request the data from the input node and count splits. -
input_types
= ['TimeSeries', 'FeatureVector', 'PredictionVector']¶
OnlyTrainSinkNode
¶
-
class
pySPACE.missions.nodes.sink.nil_sink.
OnlyTrainSinkNode
(**kwargs)[source]¶ Bases:
pySPACE.missions.nodes.sink.nil_sink.NilSinkNode
Store only meta information and perform training but not testing
The node performs only training on the node chain, so that the test procedure can be performed manually, e.g. for debug and testing reasons.
The node is very similar to the NilSinkNode.
Parameters
Exemplary Call
- node: Only_Train_Sink
Author: Hendrik Woehrle (hendrik.woehrle@dfki.de)
Created: 2011/07/14
POSSIBLE NODE NAMES: - OnlyTrainSinkNode
- OnlyTrainSink
- Only_Train_Sink
POSSIBLE INPUT TYPES: - TimeSeries
- FeatureVector
- PredictionVector
Class Components Summary
process_current_split
()Request the data from the input node and count splits request_data_for_testing
()Request data for testing, just call the predecessors method