org.deeplearning4j.arbiter.task

Interface TaskListener

All Superinterfaces:

Serializable

public interface TaskListener
extends Serializable

TaskListener: can be used to preprocess and post process a model (MultiLayerNetwork or ComputationGraph) before/after training, in a MultiLayerNetworkTaskCreator or ComputationGraphTaskCreator

Author:

Alex Black

Method Summary

Modifier and Type	Method and Description
void	postProcess(org.deeplearning4j.nn.api.Model model, Candidate candidate) Post process the model, after any training has taken place
<T extends org.deeplearning4j.nn.api.Model> T	preProcess(T model, Candidate candidate) Preprocess the model, before any training has taken place.

Method Detail

preProcess

<T extends org.deeplearning4j.nn.api.Model> T preProcess(T model,
                                                         Candidate candidate)

Preprocess the model, before any training has taken place.
Can be used to (for example) set listeners on a model before training starts

Parameters:

model - Model to preprocess

candidate - Candidate information, for the current model

Returns:

The updated model (usually the same one as the input, perhaps with modifications)

postProcess

void postProcess(org.deeplearning4j.nn.api.Model model,
                 Candidate candidate)

Post process the model, after any training has taken place

Parameters:

model - Model to postprocess

candidate - Candidate information, for the current model