Graph rewriting is experimental and subject to change, not covered by any API stability guarantees.
tensorflow.RewriterConfig
type tensorflow.RewriterConfig
Métodos Públicos
RewriterConfig.Builder | addAllCustomOptimizers (valores Iterable <? extends RewriterConfig.CustomGraphOptimizer >) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | addAllOptimizers (valores Iterable <String>) If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
RewriterConfig.Builder | addCustomOptimizers ( RewriterConfig.CustomGraphOptimizer.Builder builderForValue) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | addCustomOptimizers (int índice, RewriterConfig.CustomGraphOptimizer.Builder builderForValue) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | addCustomOptimizers (valor RewriterConfig.CustomGraphOptimizer ) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | addCustomOptimizers (int index, RewriterConfig.CustomGraphOptimizer value) list of CustomGraphOptimizers to apply. |
RewriterConfig.CustomGraphOptimizer.Builder | |
RewriterConfig.CustomGraphOptimizer.Builder | addCustomOptimizersBuilder () list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | addOptimizers (valor String) If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
RewriterConfig.Builder | addOptimizersBytes (com.google.protobuf.ByteString value) If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
RewriterConfig.Builder | addRepeatedField (campo com.google.protobuf.Descriptors.FieldDescriptor, valor do objeto) |
RewriterConfig | build () |
RewriterConfig | buildPartial () |
RewriterConfig.Builder | limpar () |
RewriterConfig.Builder | clearArithmeticOptimization () Arithmetic optimizations (default is ON) e.g. |
RewriterConfig.Builder | clearAutoMixedPrecision () Optimize data types for CUDA (default is OFF). |
RewriterConfig.Builder | clearAutoMixedPrecisionMkl () Optimize data types for MKL (default is OFF). |
RewriterConfig.Builder | clearAutoParallel () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
RewriterConfig.Builder | clearCommonSubgraphElimination () Common subgraph elimination (default is ON) e.g. |
RewriterConfig.Builder | clearConstantFolding () Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
RewriterConfig.Builder | clearCpuLayoutConversion () CPU Conversion settings between NHCW and NCHW. |
RewriterConfig.Builder | clearCustomOptimizers () list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | clearDebugStripper () Strips debug-related nodes from the graph (off by default). |
RewriterConfig.Builder | clearDependencyOptimization () Control dependency optimizations (default is ON). |
RewriterConfig.Builder | clearDisableMetaOptimizer () Disable the entire meta optimizer (off by default). |
RewriterConfig.Builder | clearDisableModelPruning () If true, don't remove unnecessary ops from the graph bool disable_model_pruning = 2; |
RewriterConfig.Builder | clearExperimentalDisableCompressedTensorOptimization () Disable optimizations that assume compressed tensors. |
RewriterConfig.Builder | clearFailOnOptimizerErrors () If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. |
RewriterConfig.Builder | clearField (campo com.google.protobuf.Descriptors.FieldDescriptor) |
RewriterConfig.Builder | clearFunctionOptimization () Function optimizations (default is ON). |
RewriterConfig.Builder | clearImplementationSelector () Enable the swap of kernel implementations based on the device placement (default is ON). |
RewriterConfig.Builder | clearInterOptimizerVerifierConfig () VerifierConfig specifying the verifiers to be run after every optimizer. |
RewriterConfig.Builder | clearLayoutOptimizer () Optimize tensor layouts (default is ON) e.g. |
RewriterConfig.Builder | clearLoopOptimization () Loop optimizations (default is ON). |
RewriterConfig.Builder | clearMemoryOptimization () Configures memory optimization passes through the meta-optimizer. |
RewriterConfig.Builder | clearMemoryOptimizerTargetNodeNameScope () A node name scope for node names which are valid outputs of recomputations. |
RewriterConfig.Builder | clearMetaOptimizerIterations () Controls how many times we run the optimizers in meta optimizer (default is once). |
RewriterConfig.Builder | clearMetaOptimizerTimeoutMs () Maximum number of milliseconds to spend optimizing a single graph before timing out. |
RewriterConfig.Builder | clearMinGraphNodes () The minimum number of nodes in a graph to optimizer. |
RewriterConfig.Builder | clearOneof (com.google.protobuf.Descriptors.OneofDescriptor oneof) |
RewriterConfig.Builder | clearOptimizers () If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
RewriterConfig.Builder | clearPinToHostOptimization () Force small ops onto the CPU (default is OFF). |
RewriterConfig.Builder | clearPostOptimizationVerifierConfig () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
RewriterConfig.Builder | clearRemapping () Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
RewriterConfig.Builder | clearScopedAllocatorOptimization () Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
RewriterConfig.Builder | clearScopedAllocatorOpts () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
RewriterConfig.Builder | clearShapeOptimization () Shape optimizations (default is ON) Simplify computations made on shapes. |
RewriterConfig.Builder | clone () |
RewriterConfig.Toggle | getArithmeticOptimization () Arithmetic optimizations (default is ON) e.g. |
int | getArithmeticOptimizationValue () Arithmetic optimizations (default is ON) e.g. |
RewriterConfig.Toggle | getAutoMixedPrecision () Optimize data types for CUDA (default is OFF). |
RewriterConfig.Toggle | getAutoMixedPrecisionMkl () Optimize data types for MKL (default is OFF). |
int | getAutoMixedPrecisionMklValue () Optimize data types for MKL (default is OFF). |
int | getAutoMixedPrecisionValue () Optimize data types for CUDA (default is OFF). |
AutoParallelOptions | getAutoParallel () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
AutoParallelOptions.Builder | getAutoParallelBuilder () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
AutoParallelOptionsOrBuilder | getAutoParallelOrBuilder () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
RewriterConfig.Toggle | getCommonSubgraphElimination () Common subgraph elimination (default is ON) e.g. |
int | getCommonSubgraphEliminationValue () Common subgraph elimination (default is ON) e.g. |
RewriterConfig.Toggle | getConstantFolding () Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
int | getConstantFoldingValue () Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
RewriterConfig.CpuLayout | getCpuLayoutConversion () CPU Conversion settings between NHCW and NCHW. |
int | getCpuLayoutConversionValue () CPU Conversion settings between NHCW and NCHW. |
RewriterConfig.CustomGraphOptimizer | getCustomOptimizers (int index) list of CustomGraphOptimizers to apply. |
RewriterConfig.CustomGraphOptimizer.Builder | getCustomOptimizersBuilder (índice int) list of CustomGraphOptimizers to apply. |
List < RewriterConfig.CustomGraphOptimizer.Builder > | getCustomOptimizersBuilderList () list of CustomGraphOptimizers to apply. |
int | getCustomOptimizersCount () list of CustomGraphOptimizers to apply. |
List < RewriterConfig.CustomGraphOptimizer > | getCustomOptimizersList () list of CustomGraphOptimizers to apply. |
RewriterConfig.CustomGraphOptimizerOrBuilder | getCustomOptimizersOrBuilder (índice int) list of CustomGraphOptimizers to apply. |
Lista <? estende RewriterConfig.CustomGraphOptimizerOrBuilder > | getCustomOptimizersOrBuilderList () list of CustomGraphOptimizers to apply. |
RewriterConfig.Toggle | getDebugStripper () Strips debug-related nodes from the graph (off by default). |
int | getDebugStripperValue () Strips debug-related nodes from the graph (off by default). |
RewriterConfig | |
RewriterConfig.Toggle | getDependencyOptimization () Control dependency optimizations (default is ON). |
int | getDependencyOptimizationValue () Control dependency optimizations (default is ON). |
final static com.google.protobuf.Descriptors.Descriptor | |
com.google.protobuf.Descriptors.Descriptor | |
boleano | getDisableMetaOptimizer () Disable the entire meta optimizer (off by default). |
boleano | getDisableModelPruning () If true, don't remove unnecessary ops from the graph bool disable_model_pruning = 2; |
boleano | getExperimentalDisableCompressedTensorOptimization () Disable optimizations that assume compressed tensors. |
boleano | getFailOnOptimizerErrors () If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. |
RewriterConfig.Toggle | getFunctionOptimization () Function optimizations (default is ON). |
int | getFunctionOptimizationValue () Function optimizations (default is ON). |
RewriterConfig.Toggle | getImplementationSelector () Enable the swap of kernel implementations based on the device placement (default is ON). |
int | getImplementationSelectorValue () Enable the swap of kernel implementations based on the device placement (default is ON). |
VerifierConfig | getInterOptimizerVerifierConfig () VerifierConfig specifying the verifiers to be run after every optimizer. |
VerifierConfig.Builder | getInterOptimizerVerifierConfigBuilder () VerifierConfig specifying the verifiers to be run after every optimizer. |
VerifierConfigOrBuilder | getInterOptimizerVerifierConfigOrBuilder () VerifierConfig specifying the verifiers to be run after every optimizer. |
RewriterConfig.Toggle | getLayoutOptimizer () Optimize tensor layouts (default is ON) e.g. |
int | getLayoutOptimizerValue () Optimize tensor layouts (default is ON) e.g. |
RewriterConfig.Toggle | getLoopOptimization () Loop optimizations (default is ON). |
int | getLoopOptimizationValue () Loop optimizations (default is ON). |
RewriterConfig.MemOptType | getMemoryOptimization () Configures memory optimization passes through the meta-optimizer. |
int | getMemoryOptimizationValue () Configures memory optimization passes through the meta-optimizer. |
Fragmento | getMemoryOptimizerTargetNodeNameScope () A node name scope for node names which are valid outputs of recomputations. |
com.google.protobuf.ByteString | getMemoryOptimizerTargetNodeNameScopeBytes () A node name scope for node names which are valid outputs of recomputations. |
RewriterConfig.NumIterationsType | getMetaOptimizerIterations () Controls how many times we run the optimizers in meta optimizer (default is once). |
int | getMetaOptimizerIterationsValue () Controls how many times we run the optimizers in meta optimizer (default is once). |
longo | getMetaOptimizerTimeoutMs () Maximum number of milliseconds to spend optimizing a single graph before timing out. |
int | getMinGraphNodes () The minimum number of nodes in a graph to optimizer. |
Fragmento | getOptimizers (int index) If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
com.google.protobuf.ByteString | getOptimizersBytes (int index) If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
int | getOptimizersCount () If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
com.google.protobuf.ProtocolStringList | getOptimizersList () If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
RewriterConfig.Toggle | getPinToHostOptimization () Force small ops onto the CPU (default is OFF). |
int | getPinToHostOptimizationValue () Force small ops onto the CPU (default is OFF). |
VerifierConfig | getPostOptimizationVerifierConfig () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
VerifierConfig.Builder | getPostOptimizationVerifierConfigBuilder () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
VerifierConfigOrBuilder | getPostOptimizationVerifierConfigOrBuilder () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
RewriterConfig.Toggle | getRemapping () Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
int | getRemappingValue () Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
RewriterConfig.Toggle | getScopedAllocatorOptimization () Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
int | getScopedAllocatorOptimizationValue () Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
ScopedAllocatorOptions | getScopedAllocatorOpts () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
ScopedAllocatorOptions.Builder | getScopedAllocatorOptsBuilder () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
ScopedAllocatorOptionsOrBuilder | getScopedAllocatorOptsOrBuilder () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
RewriterConfig.Toggle | getShapeOptimization () Shape optimizations (default is ON) Simplify computations made on shapes. |
int | getShapeOptimizationValue () Shape optimizations (default is ON) Simplify computations made on shapes. |
boleano | hasAutoParallel () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
boleano | hasInterOptimizerVerifierConfig () VerifierConfig specifying the verifiers to be run after every optimizer. |
boleano | hasPostOptimizationVerifierConfig () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
boleano | hasScopedAllocatorOpts () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
final booleano | |
RewriterConfig.Builder | mergeAutoParallel (valor AutoParallelOptions ) Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
RewriterConfig.Builder | mergeFrom (com.google.protobuf.Message other) |
RewriterConfig.Builder | mergeFrom (com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry) |
RewriterConfig.Builder | mergeInterOptimizerVerifierConfig (valor VerifierConfig ) VerifierConfig specifying the verifiers to be run after every optimizer. |
RewriterConfig.Builder | mergePostOptimizationVerifierConfig (valor VerifierConfig ) VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
RewriterConfig.Builder | mergeScopedAllocatorOpts (valor ScopedAllocatorOptions ) .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
final RewriterConfig.Builder | mergeUnknownFields (com.google.protobuf.UnknownFieldSet unknownFields) |
RewriterConfig.Builder | removeCustomOptimizers (int index) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | setArithmeticOptimization ( RewriterConfig.Toggle valor) Arithmetic optimizations (default is ON) e.g. |
RewriterConfig.Builder | setArithmeticOptimizationValue (valor int) Arithmetic optimizations (default is ON) e.g. |
RewriterConfig.Builder | |
RewriterConfig.Builder | setAutoMixedPrecisionMkl ( RewriterConfig.Toggle valor) Optimize data types for MKL (default is OFF). |
RewriterConfig.Builder | setAutoMixedPrecisionMklValue (valor int) Optimize data types for MKL (default is OFF). |
RewriterConfig.Builder | setAutoMixedPrecisionValue (valor int) Optimize data types for CUDA (default is OFF). |
RewriterConfig.Builder | setAutoParallel ( AutoParallelOptions.Builder builderForValue) Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
RewriterConfig.Builder | setAutoParallel (valor AutoParallelOptions ) Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
RewriterConfig.Builder | setCommonSubgraphElimination ( RewriterConfig.Toggle valor) Common subgraph elimination (default is ON) e.g. |
RewriterConfig.Builder | setCommonSubgraphEliminationValue (valor int) Common subgraph elimination (default is ON) e.g. |
RewriterConfig.Builder | setConstantFolding ( RewriterConfig.Toggle valor) Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
RewriterConfig.Builder | setConstantFoldingValue (valor int) Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
RewriterConfig.Builder | setCpuLayoutConversion ( RewriterConfig.CpuLayout valor) CPU Conversion settings between NHCW and NCHW. |
RewriterConfig.Builder | setCpuLayoutConversionValue (valor int) CPU Conversion settings between NHCW and NCHW. |
RewriterConfig.Builder | setCustomOptimizers (int índice, RewriterConfig.CustomGraphOptimizer.Builder builderForValue) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | setCustomOptimizers (int index, RewriterConfig.CustomGraphOptimizer value) list of CustomGraphOptimizers to apply. |
RewriterConfig.Builder | setDebugStripper ( RewriterConfig.Toggle valor) Strips debug-related nodes from the graph (off by default). |
RewriterConfig.Builder | setDebugStripperValue (valor int) Strips debug-related nodes from the graph (off by default). |
RewriterConfig.Builder | setDependencyOptimization ( RewriterConfig.Toggle valor) Control dependency optimizations (default is ON). |
RewriterConfig.Builder | setDependencyOptimizationValue (valor int) Control dependency optimizations (default is ON). |
RewriterConfig.Builder | setDisableMetaOptimizer (valor booleano) Disable the entire meta optimizer (off by default). |
RewriterConfig.Builder | setDisableModelPruning (valor booleano) If true, don't remove unnecessary ops from the graph bool disable_model_pruning = 2; |
RewriterConfig.Builder | setExperimentalDisableCompressedTensorOptimization (valor booleano) Disable optimizations that assume compressed tensors. |
RewriterConfig.Builder | setFailOnOptimizerErrors (valor booleano) If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. |
RewriterConfig.Builder | setField (campo com.google.protobuf.Descriptors.FieldDescriptor, valor do objeto) |
RewriterConfig.Builder | |
RewriterConfig.Builder | setFunctionOptimizationValue (valor int) Function optimizations (default is ON). |
RewriterConfig.Builder | setImplementationSelector ( RewriterConfig.Toggle valor) Enable the swap of kernel implementations based on the device placement (default is ON). |
RewriterConfig.Builder | setImplementationSelectorValue (valor int) Enable the swap of kernel implementations based on the device placement (default is ON). |
RewriterConfig.Builder | setInterOptimizerVerifierConfig ( VerifierConfig.Builder builderForValue) VerifierConfig specifying the verifiers to be run after every optimizer. |
RewriterConfig.Builder | setInterOptimizerVerifierConfig (valor VerifierConfig ) VerifierConfig specifying the verifiers to be run after every optimizer. |
RewriterConfig.Builder | |
RewriterConfig.Builder | setLayoutOptimizerValue (valor int) Optimize tensor layouts (default is ON) e.g. |
RewriterConfig.Builder | |
RewriterConfig.Builder | setLoopOptimizationValue (valor int) Loop optimizations (default is ON). |
RewriterConfig.Builder | setMemoryOptimization ( RewriterConfig.MemOptType valor) Configures memory optimization passes through the meta-optimizer. |
RewriterConfig.Builder | setMemoryOptimizationValue (valor int) Configures memory optimization passes through the meta-optimizer. |
RewriterConfig.Builder | setMemoryOptimizerTargetNodeNameScope (valor String) A node name scope for node names which are valid outputs of recomputations. |
RewriterConfig.Builder | setMemoryOptimizerTargetNodeNameScopeBytes (com.google.protobuf.ByteString value) A node name scope for node names which are valid outputs of recomputations. |
RewriterConfig.Builder | setMetaOptimizerIterations ( RewriterConfig.NumIterationsType valor) Controls how many times we run the optimizers in meta optimizer (default is once). |
RewriterConfig.Builder | setMetaOptimizerIterationsValue (valor int) Controls how many times we run the optimizers in meta optimizer (default is once). |
RewriterConfig.Builder | setMetaOptimizerTimeoutMs (valor longo) Maximum number of milliseconds to spend optimizing a single graph before timing out. |
RewriterConfig.Builder | setMinGraphNodes (valor int) The minimum number of nodes in a graph to optimizer. |
RewriterConfig.Builder | setOptimizers (int index, String value) If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). |
RewriterConfig.Builder | setPinToHostOptimization ( RewriterConfig.Toggle valor) Force small ops onto the CPU (default is OFF). |
RewriterConfig.Builder | setPinToHostOptimizationValue (valor int) Force small ops onto the CPU (default is OFF). |
RewriterConfig.Builder | setPostOptimizationVerifierConfig (valor VerifierConfig ) VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
RewriterConfig.Builder | setPostOptimizationVerifierConfig ( VerifierConfig.Builder builderForValue) VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
RewriterConfig.Builder | setRemapping ( RewriterConfig.Toggle valor) Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
RewriterConfig.Builder | setRemappingValue (valor int) Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
RewriterConfig.Builder | setRepeatedField (campo com.google.protobuf.Descriptors.FieldDescriptor, índice int, valor do objeto) |
RewriterConfig.Builder | setScopedAllocatorOptimization ( RewriterConfig.Toggle valor) Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
RewriterConfig.Builder | setScopedAllocatorOptimizationValue (valor int) Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
RewriterConfig.Builder | setScopedAllocatorOpts (valor ScopedAllocatorOptions ) .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
RewriterConfig.Builder | setScopedAllocatorOpts ( ScopedAllocatorOptions.Builder builderForValue) .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
RewriterConfig.Builder | setShapeOptimization ( RewriterConfig.Toggle valor) Shape optimizations (default is ON) Simplify computations made on shapes. |
RewriterConfig.Builder | setShapeOptimizationValue (valor int) Shape optimizations (default is ON) Simplify computations made on shapes. |
final RewriterConfig.Builder | setUnknownFields (com.google.protobuf.UnknownFieldSet unknownFields) |
Métodos herdados
Métodos Públicos
públicas RewriterConfig.Builder addAllCustomOptimizers (Iterable <? estende RewriterConfig.CustomGraphOptimizer > valores)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
públicas RewriterConfig.Builder addAllOptimizers (Iterable <String> valores)
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
pública RewriterConfig.Builder addCustomOptimizers ( RewriterConfig.CustomGraphOptimizer.Builder builderForValue)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
pública RewriterConfig.Builder addCustomOptimizers (int index, RewriterConfig.CustomGraphOptimizer.Builder builderForValue)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.Builder addCustomOptimizers ( RewriterConfig.CustomGraphOptimizer value)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.Builder addCustomOptimizers (int index, RewriterConfig.CustomGraphOptimizer value)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.CustomGraphOptimizer.Builder addCustomOptimizersBuilder (índice interno )
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.CustomGraphOptimizer.Builder addCustomOptimizersBuilder ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
pública RewriterConfig.Builder addOptimizers (String value)
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
pública RewriterConfig.Builder addOptimizersBytes (valor com.google.protobuf.ByteString)
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
pública RewriterConfig.Builder addRepeatedField (campo com.google.protobuf.Descriptors.FieldDescriptor, valor Object)
pública RewriterConfig.Builder clearArithmeticOptimization ()
Arithmetic optimizations (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle arithmetic_optimization = 7;
pública RewriterConfig.Builder clearAutoMixedPrecision ()
Optimize data types for CUDA (default is OFF). This will try to use float16 on GPU which is faster. Note that this can change the numerical stability of the graph and may require the use of loss scaling to maintain model convergence.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision = 23;
pública RewriterConfig.Builder clearAutoMixedPrecisionMkl ()
Optimize data types for MKL (default is OFF). This will try to use bfloat16 on CPUs, which is faster. Note that this can change the numerical stability of the graph.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision_mkl = 25;
pública RewriterConfig.Builder clearAutoParallel ()
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
pública RewriterConfig.Builder clearCommonSubgraphElimination ()
Common subgraph elimination (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle common_subgraph_elimination = 24;
pública RewriterConfig.Builder clearConstantFolding ()
Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants.
.tensorflow.RewriterConfig.Toggle constant_folding = 3;
pública RewriterConfig.Builder clearCpuLayoutConversion ()
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
pública RewriterConfig.Builder clearCustomOptimizers ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
pública RewriterConfig.Builder clearDebugStripper ()
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
pública RewriterConfig.Builder clearDependencyOptimization ()
Control dependency optimizations (default is ON). Remove redundant control dependencies, which may enable other optimization.
.tensorflow.RewriterConfig.Toggle dependency_optimization = 8;
pública RewriterConfig.Builder clearDisableMetaOptimizer ()
Disable the entire meta optimizer (off by default).
bool disable_meta_optimizer = 19;
pública RewriterConfig.Builder clearDisableModelPruning ()
If true, don't remove unnecessary ops from the graph
bool disable_model_pruning = 2;
pública RewriterConfig.Builder clearExperimentalDisableCompressedTensorOptimization ()
Disable optimizations that assume compressed tensors. Note that this flag is experimental and may be removed in the future.
bool experimental_disable_compressed_tensor_optimization = 26;
pública RewriterConfig.Builder clearFailOnOptimizerErrors ()
If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. By default - or when set to false, failing passes are skipped silently.
bool fail_on_optimizer_errors = 21;
pública RewriterConfig.Builder clearFunctionOptimization ()
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
pública RewriterConfig.Builder clearImplementationSelector ()
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
pública RewriterConfig.Builder clearInterOptimizerVerifierConfig ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
pública RewriterConfig.Builder clearLayoutOptimizer ()
Optimize tensor layouts (default is ON) e.g. This will try to use NCHW layout on GPU which is faster.
.tensorflow.RewriterConfig.Toggle layout_optimizer = 1;
pública RewriterConfig.Builder clearLoopOptimization ()
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
pública RewriterConfig.Builder clearMemoryOptimization ()
Configures memory optimization passes through the meta-optimizer. Has no effect on manually requested memory optimization passes in the optimizers field.
.tensorflow.RewriterConfig.MemOptType memory_optimization = 4;
pública RewriterConfig.Builder clearMemoryOptimizerTargetNodeNameScope ()
A node name scope for node names which are valid outputs of recomputations. Inputs to nodes that match this scope may be recomputed (subject either to manual annotation of those input nodes or to manual annotation and heuristics depending on memory_optimization), but the nodes themselves will not be recomputed. This matches any sub-scopes as well, meaning the scope can appear not just as a top-level scope. For example, if the value is "gradients/", the default, it will match node name "gradients/foo", "foo/gradients/bar", but not "foo_gradients/"
string memory_optimizer_target_node_name_scope = 6;
pública RewriterConfig.Builder clearMetaOptimizerIterations ()
Controls how many times we run the optimizers in meta optimizer (default is once).
.tensorflow.RewriterConfig.NumIterationsType meta_optimizer_iterations = 12;
pública RewriterConfig.Builder clearMetaOptimizerTimeoutMs ()
Maximum number of milliseconds to spend optimizing a single graph before timing out. If equal to 0 the system picks a default (currently 5 minutes). If less than 0 the optimizer will never time out.
int64 meta_optimizer_timeout_ms = 20;
pública RewriterConfig.Builder clearMinGraphNodes ()
The minimum number of nodes in a graph to optimizer. For smaller graphs, optimization is skipped. 0 means the system picks an appropriate number. < 0 means do not skip optimization.
int32 min_graph_nodes = 17;
pública RewriterConfig.Builder clearOptimizers ()
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
pública RewriterConfig.Builder clearPinToHostOptimization ()
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
pública RewriterConfig.Builder clearPostOptimizationVerifierConfig ()
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
pública RewriterConfig.Builder clearRemapping ()
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
pública RewriterConfig.Builder clearScopedAllocatorOptimization ()
Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default).
.tensorflow.RewriterConfig.Toggle scoped_allocator_optimization = 15;
pública RewriterConfig.Builder clearScopedAllocatorOpts ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
pública RewriterConfig.Builder clearShapeOptimization ()
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;
pública RewriterConfig.Toggle getArithmeticOptimization ()
Arithmetic optimizations (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle arithmetic_optimization = 7;
public int getArithmeticOptimizationValue ()
Arithmetic optimizations (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle arithmetic_optimization = 7;
pública RewriterConfig.Toggle getAutoMixedPrecision ()
Optimize data types for CUDA (default is OFF). This will try to use float16 on GPU which is faster. Note that this can change the numerical stability of the graph and may require the use of loss scaling to maintain model convergence.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision = 23;
pública RewriterConfig.Toggle getAutoMixedPrecisionMkl ()
Optimize data types for MKL (default is OFF). This will try to use bfloat16 on CPUs, which is faster. Note that this can change the numerical stability of the graph.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision_mkl = 25;
public int getAutoMixedPrecisionMklValue ()
Optimize data types for MKL (default is OFF). This will try to use bfloat16 on CPUs, which is faster. Note that this can change the numerical stability of the graph.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision_mkl = 25;
public int getAutoMixedPrecisionValue ()
Optimize data types for CUDA (default is OFF). This will try to use float16 on GPU which is faster. Note that this can change the numerical stability of the graph and may require the use of loss scaling to maintain model convergence.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision = 23;
public AutoParallelOptions getAutoParallel ()
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
public AutoParallelOptions.Builder getAutoParallelBuilder ()
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
public AutoParallelOptionsOrBuilder getAutoParallelOrBuilder ()
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
pública RewriterConfig.Toggle getCommonSubgraphElimination ()
Common subgraph elimination (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle common_subgraph_elimination = 24;
public int getCommonSubgraphEliminationValue ()
Common subgraph elimination (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle common_subgraph_elimination = 24;
pública RewriterConfig.Toggle getConstantFolding ()
Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants.
.tensorflow.RewriterConfig.Toggle constant_folding = 3;
public int getConstantFoldingValue ()
Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants.
.tensorflow.RewriterConfig.Toggle constant_folding = 3;
pública RewriterConfig.CpuLayout getCpuLayoutConversion ()
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
public int getCpuLayoutConversionValue ()
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
pública RewriterConfig.CustomGraphOptimizer getCustomOptimizers (int index)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
pública RewriterConfig.CustomGraphOptimizer.Builder getCustomOptimizersBuilder (int index)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
Lista pública < RewriterConfig.CustomGraphOptimizer.Builder > getCustomOptimizersBuilderList ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public int getCustomOptimizersCount ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
Lista pública < RewriterConfig.CustomGraphOptimizer > getCustomOptimizersList ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
pública RewriterConfig.CustomGraphOptimizerOrBuilder getCustomOptimizersOrBuilder (int index)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
Lista pública <? estende RewriterConfig.CustomGraphOptimizerOrBuilder > getCustomOptimizersOrBuilderList ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
pública RewriterConfig.Toggle getDebugStripper ()
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
public int getDebugStripperValue ()
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
pública RewriterConfig.Toggle getDependencyOptimization ()
Control dependency optimizations (default is ON). Remove redundant control dependencies, which may enable other optimization.
.tensorflow.RewriterConfig.Toggle dependency_optimization = 8;
public int getDependencyOptimizationValue ()
Control dependency optimizations (default is ON). Remove redundant control dependencies, which may enable other optimization.
.tensorflow.RewriterConfig.Toggle dependency_optimization = 8;
public static final com.google.protobuf.Descriptors.Descriptor getDescriptor ()
public com.google.protobuf.Descriptors.Descriptor getDescriptorForType ()
public boolean getDisableMetaOptimizer ()
Disable the entire meta optimizer (off by default).
bool disable_meta_optimizer = 19;
public boolean getDisableModelPruning ()
If true, don't remove unnecessary ops from the graph
bool disable_model_pruning = 2;
public boolean getExperimentalDisableCompressedTensorOptimization ()
Disable optimizations that assume compressed tensors. Note that this flag is experimental and may be removed in the future.
bool experimental_disable_compressed_tensor_optimization = 26;
public boolean getFailOnOptimizerErrors ()
If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. By default - or when set to false, failing passes are skipped silently.
bool fail_on_optimizer_errors = 21;
pública RewriterConfig.Toggle getFunctionOptimization ()
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
public int getFunctionOptimizationValue ()
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
pública RewriterConfig.Toggle getImplementationSelector ()
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
public int getImplementationSelectorValue ()
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
public VerifierConfig getInterOptimizerVerifierConfig ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public VerifierConfig.Builder getInterOptimizerVerifierConfigBuilder ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public VerifierConfigOrBuilder getInterOptimizerVerifierConfigOrBuilder ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
pública RewriterConfig.Toggle getLayoutOptimizer ()
Optimize tensor layouts (default is ON) e.g. This will try to use NCHW layout on GPU which is faster.
.tensorflow.RewriterConfig.Toggle layout_optimizer = 1;
public int getLayoutOptimizerValue ()
Optimize tensor layouts (default is ON) e.g. This will try to use NCHW layout on GPU which is faster.
.tensorflow.RewriterConfig.Toggle layout_optimizer = 1;
pública RewriterConfig.Toggle getLoopOptimization ()
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
public int getLoopOptimizationValue ()
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
pública RewriterConfig.MemOptType getMemoryOptimization ()
Configures memory optimization passes through the meta-optimizer. Has no effect on manually requested memory optimization passes in the optimizers field.
.tensorflow.RewriterConfig.MemOptType memory_optimization = 4;
public int getMemoryOptimizationValue ()
Configures memory optimization passes through the meta-optimizer. Has no effect on manually requested memory optimization passes in the optimizers field.
.tensorflow.RewriterConfig.MemOptType memory_optimization = 4;
public String getMemoryOptimizerTargetNodeNameScope ()
A node name scope for node names which are valid outputs of recomputations. Inputs to nodes that match this scope may be recomputed (subject either to manual annotation of those input nodes or to manual annotation and heuristics depending on memory_optimization), but the nodes themselves will not be recomputed. This matches any sub-scopes as well, meaning the scope can appear not just as a top-level scope. For example, if the value is "gradients/", the default, it will match node name "gradients/foo", "foo/gradients/bar", but not "foo_gradients/"
string memory_optimizer_target_node_name_scope = 6;
public com.google.protobuf.ByteString getMemoryOptimizerTargetNodeNameScopeBytes ()
A node name scope for node names which are valid outputs of recomputations. Inputs to nodes that match this scope may be recomputed (subject either to manual annotation of those input nodes or to manual annotation and heuristics depending on memory_optimization), but the nodes themselves will not be recomputed. This matches any sub-scopes as well, meaning the scope can appear not just as a top-level scope. For example, if the value is "gradients/", the default, it will match node name "gradients/foo", "foo/gradients/bar", but not "foo_gradients/"
string memory_optimizer_target_node_name_scope = 6;
pública RewriterConfig.NumIterationsType getMetaOptimizerIterations ()
Controls how many times we run the optimizers in meta optimizer (default is once).
.tensorflow.RewriterConfig.NumIterationsType meta_optimizer_iterations = 12;
public int getMetaOptimizerIterationsValue ()
Controls how many times we run the optimizers in meta optimizer (default is once).
.tensorflow.RewriterConfig.NumIterationsType meta_optimizer_iterations = 12;
public long getMetaOptimizerTimeoutMs ()
Maximum number of milliseconds to spend optimizing a single graph before timing out. If equal to 0 the system picks a default (currently 5 minutes). If less than 0 the optimizer will never time out.
int64 meta_optimizer_timeout_ms = 20;
public int getMinGraphNodes ()
The minimum number of nodes in a graph to optimizer. For smaller graphs, optimization is skipped. 0 means the system picks an appropriate number. < 0 means do not skip optimization.
int32 min_graph_nodes = 17;
public String getOptimizers (int index)
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
public com.google.protobuf.ByteString getOptimizersBytes (int index)
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
public int getOptimizersCount ()
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
public com.google.protobuf.ProtocolStringList getOptimizersList ()
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
pública RewriterConfig.Toggle getPinToHostOptimization ()
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
public int getPinToHostOptimizationValue ()
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
public VerifierConfig getPostOptimizationVerifierConfig ()
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public VerifierConfig.Builder getPostOptimizationVerifierConfigBuilder ()
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public VerifierConfigOrBuilder getPostOptimizationVerifierConfigOrBuilder ()
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
pública RewriterConfig.Toggle getRemapping ()
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
public int getRemappingValue ()
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
pública RewriterConfig.Toggle getScopedAllocatorOptimization ()
Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default).
.tensorflow.RewriterConfig.Toggle scoped_allocator_optimization = 15;
public int getScopedAllocatorOptimizationValue ()
Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default).
.tensorflow.RewriterConfig.Toggle scoped_allocator_optimization = 15;
public ScopedAllocatorOptions getScopedAllocatorOpts ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public ScopedAllocatorOptions.Builder getScopedAllocatorOptsBuilder ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public ScopedAllocatorOptionsOrBuilder getScopedAllocatorOptsOrBuilder ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
pública RewriterConfig.Toggle getShapeOptimization ()
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;
public int getShapeOptimizationValue ()
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;
public boolean hasAutoParallel ()
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
public boolean hasInterOptimizerVerifierConfig ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public boolean hasPostOptimizationVerifierConfig ()
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public boolean hasScopedAllocatorOpts ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public final boolean isInitialized ()
pública RewriterConfig.Builder mergeAutoParallel ( AutoParallelOptions valor)
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
pública RewriterConfig.Builder mergefrom (entrada com.google.protobuf.CodedInputStream, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
Lança
IOException |
---|
pública RewriterConfig.Builder mergeInterOptimizerVerifierConfig ( VerifierConfig valor)
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
pública RewriterConfig.Builder mergePostOptimizationVerifierConfig ( VerifierConfig valor)
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public RewriterConfig.Builder mergeScopedAllocatorOpts ( ScopedAllocatorOptions value)
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public final RewriterConfig.Builder mergeUnknownFields (com.google.protobuf.UnknownFieldSet unknownFields)
public RewriterConfig.Builder removeCustomOptimizers (int index)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.Builder setArithmeticOptimization ( RewriterConfig.Toggle value)
Arithmetic optimizations (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle arithmetic_optimization = 7;
public RewriterConfig.Builder setArithmeticOptimizationValue (int value)
Arithmetic optimizations (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle arithmetic_optimization = 7;
public RewriterConfig.Builder setAutoMixedPrecision ( RewriterConfig.Toggle value)
Optimize data types for CUDA (default is OFF). This will try to use float16 on GPU which is faster. Note that this can change the numerical stability of the graph and may require the use of loss scaling to maintain model convergence.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision = 23;
public RewriterConfig.Builder setAutoMixedPrecisionMkl ( RewriterConfig.Toggle value)
Optimize data types for MKL (default is OFF). This will try to use bfloat16 on CPUs, which is faster. Note that this can change the numerical stability of the graph.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision_mkl = 25;
public RewriterConfig.Builder setAutoMixedPrecisionMklValue (int value)
Optimize data types for MKL (default is OFF). This will try to use bfloat16 on CPUs, which is faster. Note that this can change the numerical stability of the graph.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision_mkl = 25;
public RewriterConfig.Builder setAutoMixedPrecisionValue (int value)
Optimize data types for CUDA (default is OFF). This will try to use float16 on GPU which is faster. Note that this can change the numerical stability of the graph and may require the use of loss scaling to maintain model convergence.
.tensorflow.RewriterConfig.Toggle auto_mixed_precision = 23;
public RewriterConfig.Builder setAutoParallel ( AutoParallelOptions.Builder builderForValue)
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
public RewriterConfig.Builder setAutoParallel ( AutoParallelOptions value)
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
public RewriterConfig.Builder setCommonSubgraphElimination ( RewriterConfig.Toggle value)
Common subgraph elimination (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle common_subgraph_elimination = 24;
public RewriterConfig.Builder setCommonSubgraphEliminationValue (int value)
Common subgraph elimination (default is ON) e.g. Simplify arithmetic ops; merge ops with same value (like constants).
.tensorflow.RewriterConfig.Toggle common_subgraph_elimination = 24;
public RewriterConfig.Builder setConstantFolding ( RewriterConfig.Toggle value)
Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants.
.tensorflow.RewriterConfig.Toggle constant_folding = 3;
public RewriterConfig.Builder setConstantFoldingValue (int value)
Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants.
.tensorflow.RewriterConfig.Toggle constant_folding = 3;
public RewriterConfig.Builder setCpuLayoutConversion ( RewriterConfig.CpuLayout value)
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
public RewriterConfig.Builder setCpuLayoutConversionValue (int value)
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
public RewriterConfig.Builder setCustomOptimizers (int index, RewriterConfig.CustomGraphOptimizer.Builder builderForValue)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.Builder setCustomOptimizers (int index, RewriterConfig.CustomGraphOptimizer value)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public RewriterConfig.Builder setDebugStripper ( RewriterConfig.Toggle value)
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
public RewriterConfig.Builder setDebugStripperValue (int value)
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
public RewriterConfig.Builder setDependencyOptimization ( RewriterConfig.Toggle value)
Control dependency optimizations (default is ON). Remove redundant control dependencies, which may enable other optimization.
.tensorflow.RewriterConfig.Toggle dependency_optimization = 8;
public RewriterConfig.Builder setDependencyOptimizationValue (int value)
Control dependency optimizations (default is ON). Remove redundant control dependencies, which may enable other optimization.
.tensorflow.RewriterConfig.Toggle dependency_optimization = 8;
public RewriterConfig.Builder setDisableMetaOptimizer (boolean value)
Disable the entire meta optimizer (off by default).
bool disable_meta_optimizer = 19;
public RewriterConfig.Builder setDisableModelPruning (boolean value)
If true, don't remove unnecessary ops from the graph
bool disable_model_pruning = 2;
public RewriterConfig.Builder setExperimentalDisableCompressedTensorOptimization (boolean value)
Disable optimizations that assume compressed tensors. Note that this flag is experimental and may be removed in the future.
bool experimental_disable_compressed_tensor_optimization = 26;
public RewriterConfig.Builder setFailOnOptimizerErrors (boolean value)
If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. By default - or when set to false, failing passes are skipped silently.
bool fail_on_optimizer_errors = 21;
public RewriterConfig.Builder setField (com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
public RewriterConfig.Builder setFunctionOptimization ( RewriterConfig.Toggle value)
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
public RewriterConfig.Builder setFunctionOptimizationValue (int value)
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
public RewriterConfig.Builder setImplementationSelector ( RewriterConfig.Toggle value)
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
public RewriterConfig.Builder setImplementationSelectorValue (int value)
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
public RewriterConfig.Builder setInterOptimizerVerifierConfig ( VerifierConfig.Builder builderForValue)
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public RewriterConfig.Builder setInterOptimizerVerifierConfig ( VerifierConfig value)
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public RewriterConfig.Builder setLayoutOptimizer ( RewriterConfig.Toggle value)
Optimize tensor layouts (default is ON) e.g. This will try to use NCHW layout on GPU which is faster.
.tensorflow.RewriterConfig.Toggle layout_optimizer = 1;
public RewriterConfig.Builder setLayoutOptimizerValue (int value)
Optimize tensor layouts (default is ON) e.g. This will try to use NCHW layout on GPU which is faster.
.tensorflow.RewriterConfig.Toggle layout_optimizer = 1;
public RewriterConfig.Builder setLoopOptimization ( RewriterConfig.Toggle value)
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
public RewriterConfig.Builder setLoopOptimizationValue (int value)
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
public RewriterConfig.Builder setMemoryOptimization ( RewriterConfig.MemOptType value)
Configures memory optimization passes through the meta-optimizer. Has no effect on manually requested memory optimization passes in the optimizers field.
.tensorflow.RewriterConfig.MemOptType memory_optimization = 4;
public RewriterConfig.Builder setMemoryOptimizationValue (int value)
Configures memory optimization passes through the meta-optimizer. Has no effect on manually requested memory optimization passes in the optimizers field.
.tensorflow.RewriterConfig.MemOptType memory_optimization = 4;
public RewriterConfig.Builder setMemoryOptimizerTargetNodeNameScope (String value)
A node name scope for node names which are valid outputs of recomputations. Inputs to nodes that match this scope may be recomputed (subject either to manual annotation of those input nodes or to manual annotation and heuristics depending on memory_optimization), but the nodes themselves will not be recomputed. This matches any sub-scopes as well, meaning the scope can appear not just as a top-level scope. For example, if the value is "gradients/", the default, it will match node name "gradients/foo", "foo/gradients/bar", but not "foo_gradients/"
string memory_optimizer_target_node_name_scope = 6;
public RewriterConfig.Builder setMemoryOptimizerTargetNodeNameScopeBytes (com.google.protobuf.ByteString value)
A node name scope for node names which are valid outputs of recomputations. Inputs to nodes that match this scope may be recomputed (subject either to manual annotation of those input nodes or to manual annotation and heuristics depending on memory_optimization), but the nodes themselves will not be recomputed. This matches any sub-scopes as well, meaning the scope can appear not just as a top-level scope. For example, if the value is "gradients/", the default, it will match node name "gradients/foo", "foo/gradients/bar", but not "foo_gradients/"
string memory_optimizer_target_node_name_scope = 6;
public RewriterConfig.Builder setMetaOptimizerIterations ( RewriterConfig.NumIterationsType value)
Controls how many times we run the optimizers in meta optimizer (default is once).
.tensorflow.RewriterConfig.NumIterationsType meta_optimizer_iterations = 12;
public RewriterConfig.Builder setMetaOptimizerIterationsValue (int value)
Controls how many times we run the optimizers in meta optimizer (default is once).
.tensorflow.RewriterConfig.NumIterationsType meta_optimizer_iterations = 12;
public RewriterConfig.Builder setMetaOptimizerTimeoutMs (long value)
Maximum number of milliseconds to spend optimizing a single graph before timing out. If equal to 0 the system picks a default (currently 5 minutes). If less than 0 the optimizer will never time out.
int64 meta_optimizer_timeout_ms = 20;
public RewriterConfig.Builder setMinGraphNodes (int value)
The minimum number of nodes in a graph to optimizer. For smaller graphs, optimization is skipped. 0 means the system picks an appropriate number. < 0 means do not skip optimization.
int32 min_graph_nodes = 17;
public RewriterConfig.Builder setOptimizers (int index, String value)
If non-empty, will use this as an alternative way to specify a list of optimizations to turn on and the order of the optimizations (replacing the meta-optimizer). Of the RewriterConfig options, only the AutoParallel configuration options (the auto_parallel field) apply to manually requested optimization passes ("autoparallel"). Memory optimization passes ("memory") invoked here are not configurable (in contrast to memory optimization passes through the meta-optimizer) and act only on manual op annotations. Custom optimizers (see custom_optimizers) that are not part of this schedule will be run after - in the order that they were specified.
repeated string optimizers = 100;
public RewriterConfig.Builder setPinToHostOptimization ( RewriterConfig.Toggle value)
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
public RewriterConfig.Builder setPinToHostOptimizationValue (int value)
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
public RewriterConfig.Builder setPostOptimizationVerifierConfig ( VerifierConfig value)
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public RewriterConfig.Builder setPostOptimizationVerifierConfig ( VerifierConfig.Builder builderForValue)
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public RewriterConfig.Builder setRemapping ( RewriterConfig.Toggle value)
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
public RewriterConfig.Builder setRemappingValue (int value)
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
public RewriterConfig.Builder setRepeatedField (com.google.protobuf.Descriptors.FieldDescriptor field, int index, Object value)
public RewriterConfig.Builder setScopedAllocatorOptimization ( RewriterConfig.Toggle value)
Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default).
.tensorflow.RewriterConfig.Toggle scoped_allocator_optimization = 15;
public RewriterConfig.Builder setScopedAllocatorOptimizationValue (int value)
Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default).
.tensorflow.RewriterConfig.Toggle scoped_allocator_optimization = 15;
public RewriterConfig.Builder setScopedAllocatorOpts ( ScopedAllocatorOptions value)
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public RewriterConfig.Builder setScopedAllocatorOpts ( ScopedAllocatorOptions.Builder builderForValue)
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public RewriterConfig.Builder setShapeOptimization ( RewriterConfig.Toggle value)
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;
public RewriterConfig.Builder setShapeOptimizationValue (int value)
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;