org.nd4j.autodiff.samediff.internal

Class InferenceSession

java.lang.Object

org.nd4j.autodiff.samediff.internal.AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

org.nd4j.autodiff.samediff.internal.InferenceSession

Direct Known Subclasses:

TrainingSession

public class InferenceSession
extends AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected static class	InferenceSession.ConstantDep
static class	InferenceSession.Dep
protected static class	InferenceSession.ExecDoneDep
static class	InferenceSession.OpDep
protected static class	InferenceSession.PlaceholderDep
protected static class	InferenceSession.ReqOutputDep
protected static class	InferenceSession.VariableDep

Nested classes/interfaces inherited from class org.nd4j.autodiff.samediff.internal.AbstractSession

AbstractSession.ExecStep, AbstractSession.ExecStepPredicate, AbstractSession.ExecType, AbstractSession.VarId

Field Summary

Fields

Modifier and Type	Field and Description
protected static String	KERAS_TRAIN_TEST

Fields inherited from class org.nd4j.autodiff.samediff.internal.AbstractSession

dt, nodeOutputs, OUTER_FRAME, sameDiff, subgraph, subgraphOps, tensorArrays, zeroInputOpsInSubgraph

Constructor Summary

Constructors

Constructor and Description
InferenceSession(@NonNull SameDiff sameDiff)

Method Summary

Modifier and Type	Method and Description
INDArray[]	doExec(DifferentialFunction op, OpContext opContext, FrameIter outputFrameIter, Set<AbstractSession.VarId> opInputs, Set<AbstractSession.VarId> allIterInputs, Set<String> constAndPhInputs)
Pair<SameDiffOp,OpContext>	getAndParameterizeOp(String opName, FrameIter frameIter, Set<AbstractSession.VarId> opInputs, Set<AbstractSession.VarId> allIterInputs, Set<String> constAndPhInputs, Map<String,INDArray> placeholderValues, Set<String> allReqVariables) Get the parameterized op to execute - for example, the op/DifferentialFunction with all inputs set
protected INDArray	getArray(SDVariable sdv, Collection<AbstractSession.VarId> opInputs, Collection<AbstractSession.VarId> allIterInputs)
INDArray	getConstantOrVariable(String variableName) Get the constant or variable output - for example, constant array or constant shape.
INDArray[]	getOutputs(Pair<SameDiffOp,OpContext> opPair, FrameIter outputFrameIter, Set<AbstractSession.VarId> opInputs, Set<AbstractSession.VarId> allIterInputs, Set<String> constAndPhInputs, List<Listener> listeners, At at, MultiDataSet batch, Set<String> allReqVariables) Execute the op - calculate INDArrays, or shape info, etc
INDArray[]	getOutputsHelperTensorArrayOps(DifferentialFunction op, FrameIter outputFrameIter, Set<AbstractSession.VarId> opInputs, Set<AbstractSession.VarId> allIterInputs) Forward pass for TensorArray ops
protected Map<String,INDArray>	postProcessOutput(Map<String,INDArray> output) Post process the session output values, if required.
protected Map<String,INDArray>	preprocessPlaceholders(Map<String,INDArray> placeholders, At at) Preprocess the placeholder values, if required.

Methods inherited from class org.nd4j.autodiff.samediff.internal.AbstractSession

addDependenciesForOp, addVarControlDeps, contains, execFailed, get, get, getExecStepForVar, initSubgraph, lookup, lookup, output, updateDescendantDeps

Methods inherited from class java.lang.Object

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

Field Detail

KERAS_TRAIN_TEST

protected static final String KERAS_TRAIN_TEST

See Also:

Constant Field Values

Constructor Detail

InferenceSession

public InferenceSession(@NonNull
                        @NonNull SameDiff sameDiff)

Method Detail

preprocessPlaceholders

protected Map<String,INDArray> preprocessPlaceholders(Map<String,INDArray> placeholders,
                                                      At at)

Description copied from class: AbstractSession

Preprocess the placeholder values, if required. Mainly reserved for casting in the case of InferenceSession

Overrides:

preprocessPlaceholders in class AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

Parameters:

placeholders - Placeholders to preprocess.

Returns:

Preprocessed placeholders

postProcessOutput

protected Map<String,INDArray> postProcessOutput(Map<String,INDArray> output)

Description copied from class: AbstractSession

Post process the session output values, if required. Override if required in session subclasses

Overrides:

postProcessOutput in class AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

Parameters:

output - Output to be returned to the user

Returns:

Post processed output

getOutputs

public INDArray[] getOutputs(Pair<SameDiffOp,OpContext> opPair,
                             FrameIter outputFrameIter,
                             Set<AbstractSession.VarId> opInputs,
                             Set<AbstractSession.VarId> allIterInputs,
                             Set<String> constAndPhInputs,
                             List<Listener> listeners,
                             At at,
                             MultiDataSet batch,
                             Set<String> allReqVariables)

Description copied from class: AbstractSession

Execute the op - calculate INDArrays, or shape info, etc

Specified by:

getOutputs in class AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

Parameters:

opPair - Operation to exit. This should be parameterized (i.e., all inputs set)

outputFrameIter - The frame and iteration of the outputs

opInputs - The specific input arrays for the op

allReqVariables - All required variables requested for the current session execution (not just the current op outputs)

Returns:

The outputs of the op

doExec

public INDArray[] doExec(DifferentialFunction op,
                         OpContext opContext,
                         FrameIter outputFrameIter,
                         Set<AbstractSession.VarId> opInputs,
                         Set<AbstractSession.VarId> allIterInputs,
                         Set<String> constAndPhInputs)

getOutputsHelperTensorArrayOps

public INDArray[] getOutputsHelperTensorArrayOps(DifferentialFunction op,
                                                 FrameIter outputFrameIter,
                                                 Set<AbstractSession.VarId> opInputs,
                                                 Set<AbstractSession.VarId> allIterInputs)

Forward pass for TensorArray ops

getConstantOrVariable

public INDArray getConstantOrVariable(String variableName)

Description copied from class: AbstractSession

Get the constant or variable output - for example, constant array or constant shape. Note that both constants and variables (i.e., VariableType.CONSTANT and VariableType.VARIABLE) are the same for all frames and iterations.

Specified by:

getConstantOrVariable in class AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

Parameters:

variableName - The name of the variable to get the constant for

Returns:

The constant

getAndParameterizeOp

public Pair<SameDiffOp,OpContext> getAndParameterizeOp(String opName,
                                                       FrameIter frameIter,
                                                       Set<AbstractSession.VarId> opInputs,
                                                       Set<AbstractSession.VarId> allIterInputs,
                                                       Set<String> constAndPhInputs,
                                                       Map<String,INDArray> placeholderValues,
                                                       Set<String> allReqVariables)

Description copied from class: AbstractSession

Get the parameterized op to execute - for example, the op/DifferentialFunction with all inputs set

Specified by:

getAndParameterizeOp in class AbstractSession<INDArray,Pair<SameDiffOp,OpContext>>

Parameters:

opName - Name of the op

frameIter - The frame and iteration of the op outputs

opInputs - The inputs to the op (excluding constants/placeholders) - for the specific frame + iteration

allIterInputs - The inputs - those that are not iteration-specific (mainly Enter op vars, which might be used in all iterations but are only executed once on iter 0)

constAndPhInputs - The constant and placeholder inputs - used for all frames/iterations

allReqVariables - All required variables requested for the current session execution (not just the current op outputs)

Returns:

The parameterized op

getArray

protected INDArray getArray(SDVariable sdv,
                            Collection<AbstractSession.VarId> opInputs,
                            Collection<AbstractSession.VarId> allIterInputs)