interface pública RewriterConfigOrBuilder
Subclasses indiretas conhecidas |
Métodos Públicos
abstrato RewriterConfig.Toggle | getArithmeticOptimization () Arithmetic optimizations (default is ON) e.g. |
int abstrato | getArithmeticOptimizationValue () Arithmetic optimizations (default is ON) e.g. |
abstrato RewriterConfig.Toggle | getAutoMixedPrecision () Optimize data types for CUDA (default is OFF). |
abstrato RewriterConfig.Toggle | getAutoMixedPrecisionMkl () Optimize data types for MKL (default is OFF). |
int abstrato | getAutoMixedPrecisionMklValue () Optimize data types for MKL (default is OFF). |
int abstrato | getAutoMixedPrecisionValue () Optimize data types for CUDA (default is OFF). |
AutoParallelOptions abstratos | getAutoParallel () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
abstrato AutoParallelOptionsOrBuilder | getAutoParallelOrBuilder () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
abstrato RewriterConfig.Toggle | getCommonSubgraphElimination () Common subgraph elimination (default is ON) e.g. |
int abstrato | getCommonSubgraphEliminationValue () Common subgraph elimination (default is ON) e.g. |
abstrato RewriterConfig.Toggle | getConstantFolding () Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
int abstrato | getConstantFoldingValue () Fold constants (default is ON) Statically infer the value of tensors when possible, and materialize the result using constants. |
abstrato RewriterConfig.CpuLayout | getCpuLayoutConversion () CPU Conversion settings between NHCW and NCHW. |
int abstrato | getCpuLayoutConversionValue () CPU Conversion settings between NHCW and NCHW. |
abstrato RewriterConfig.CustomGraphOptimizer | getCustomOptimizers (int index) list of CustomGraphOptimizers to apply. |
int abstrato | getCustomOptimizersCount () list of CustomGraphOptimizers to apply. |
Lista abstrata < RewriterConfig.CustomGraphOptimizer > | getCustomOptimizersList () list of CustomGraphOptimizers to apply. |
abstrato RewriterConfig.CustomGraphOptimizerOrBuilder | getCustomOptimizersOrBuilder (índice int) list of CustomGraphOptimizers to apply. |
Lista abstrata <? estende RewriterConfig.CustomGraphOptimizerOrBuilder > | getCustomOptimizersOrBuilderList () list of CustomGraphOptimizers to apply. |
abstrato RewriterConfig.Toggle | getDebugStripper () Strips debug-related nodes from the graph (off by default). |
int abstrato | getDebugStripperValue () Strips debug-related nodes from the graph (off by default). |
abstrato RewriterConfig.Toggle | getDependencyOptimization () Control dependency optimizations (default is ON). |
int abstrato | getDependencyOptimizationValue () Control dependency optimizations (default is ON). |
booleano abstrato | getDisableMetaOptimizer () Disable the entire meta optimizer (off by default). |
booleano abstrato | getDisableModelPruning () If true, don't remove unnecessary ops from the graph bool disable_model_pruning = 2; |
booleano abstrato | getExperimentalDisableCompressedTensorOptimization () Disable optimizations that assume compressed tensors. |
booleano abstrato | getFailOnOptimizerErrors () If true, any optimization pass failing will cause the MetaOptimizer to stop with an error. |
abstrato RewriterConfig.Toggle | getFunctionOptimization () Function optimizations (default is ON). |
int abstrato | getFunctionOptimizationValue () Function optimizations (default is ON). |
abstrato RewriterConfig.Toggle | getImplementationSelector () Enable the swap of kernel implementations based on the device placement (default is ON). |
int abstrato | getImplementationSelectorValue () Enable the swap of kernel implementations based on the device placement (default is ON). |
Resumo VerifierConfig | getInterOptimizerVerifierConfig () VerifierConfig specifying the verifiers to be run after every optimizer. |
abstrato VerifierConfigOrBuilder | getInterOptimizerVerifierConfigOrBuilder () VerifierConfig specifying the verifiers to be run after every optimizer. |
abstrato RewriterConfig.Toggle | getLayoutOptimizer () Optimize tensor layouts (default is ON) e.g. |
int abstrato | getLayoutOptimizerValue () Optimize tensor layouts (default is ON) e.g. |
abstrato RewriterConfig.Toggle | getLoopOptimization () Loop optimizations (default is ON). |
int abstrato | getLoopOptimizationValue () Loop optimizations (default is ON). |
abstrato RewriterConfig.MemOptType | getMemoryOptimization () Configures memory optimization passes through the meta-optimizer. |
int abstrato | getMemoryOptimizationValue () Configures memory optimization passes through the meta-optimizer. |
String abstrata | getMemoryOptimizerTargetNodeNameScope () A node name scope for node names which are valid outputs of recomputations. |
abstract com.google.protobuf.ByteString | getMemoryOptimizerTargetNodeNameScopeBytes () A node name scope for node names which are valid outputs of recomputations. |
abstrato RewriterConfig.NumIterationsType | getMetaOptimizerIterations () Controls how many times we run the optimizers in meta optimizer (default is once). |
int abstrato | getMetaOptimizerIterationsValue () Controls how many times we run the optimizers in meta optimizer (default is once). |
abstrato longo | getMetaOptimizerTimeoutMs () Maximum number of milliseconds to spend optimizing a single graph before timing out. |
int abstrato | getMinGraphNodes () The minimum number of nodes in a graph to optimizer. |
String abstrata | 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). |
abstract 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 abstrato | 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). |
Lista abstrata <String> | 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). |
abstrato RewriterConfig.Toggle | getPinToHostOptimization () Force small ops onto the CPU (default is OFF). |
int abstrato | getPinToHostOptimizationValue () Force small ops onto the CPU (default is OFF). |
Resumo VerifierConfig | getPostOptimizationVerifierConfig () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
abstrato VerifierConfigOrBuilder | getPostOptimizationVerifierConfigOrBuilder () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
abstrato RewriterConfig.Toggle | getRemapping () Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
int abstrato | getRemappingValue () Remapping (default is ON) Remap subgraphs onto more efficient implementations. |
abstrato RewriterConfig.Toggle | getScopedAllocatorOptimization () Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
int abstrato | getScopedAllocatorOptimizationValue () Try to allocate some independent Op outputs contiguously in order to merge or eliminate downstream Ops (off by default). |
ScopedAllocatorOptions abstrato | getScopedAllocatorOpts () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
abstrato ScopedAllocatorOptionsOrBuilder | getScopedAllocatorOptsOrBuilder () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
abstrato RewriterConfig.Toggle | getShapeOptimization () Shape optimizations (default is ON) Simplify computations made on shapes. |
int abstrato | getShapeOptimizationValue () Shape optimizations (default is ON) Simplify computations made on shapes. |
booleano abstrato | hasAutoParallel () Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field. |
booleano abstrato | hasInterOptimizerVerifierConfig () VerifierConfig specifying the verifiers to be run after every optimizer. |
booleano abstrato | hasPostOptimizationVerifierConfig () VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run. |
booleano abstrato | hasScopedAllocatorOpts () .tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16; |
Métodos Públicos
public abstract 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 abstract 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;
public abstract 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;
public abstract 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 abstract 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 abstract 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 abstract 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 abstract AutoParallelOptionsOrBuilder getAutoParallelOrBuilder ()
Configures AutoParallel optimization passes either through the meta-optimizer or when manually specified through the optimizers field.
.tensorflow.AutoParallelOptions auto_parallel = 5;
public abstract 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 abstract 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;
public abstract 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 abstract 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;
public abstract RewriterConfig.CpuLayout getCpuLayoutConversion ()
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
public abstract int getCpuLayoutConversionValue ()
CPU Conversion settings between NHCW and NCHW.
.tensorflow.RewriterConfig.CpuLayout cpu_layout_conversion = 50;
abstratos públicas RewriterConfig.CustomGraphOptimizer getCustomOptimizers (int index)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public abstract int getCustomOptimizersCount ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
Lista public abstract < RewriterConfig.CustomGraphOptimizer > getCustomOptimizersList ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
abstrato pública RewriterConfig.CustomGraphOptimizerOrBuilder getCustomOptimizersOrBuilder (int index)
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
Lista de resumos pública <? estende RewriterConfig.CustomGraphOptimizerOrBuilder > getCustomOptimizersOrBuilderList ()
list of CustomGraphOptimizers to apply.
repeated .tensorflow.RewriterConfig.CustomGraphOptimizer custom_optimizers = 200;
public abstract RewriterConfig.Toggle getDebugStripper ()
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
public abstract int getDebugStripperValue ()
Strips debug-related nodes from the graph (off by default).
.tensorflow.RewriterConfig.Toggle debug_stripper = 11;
public abstract 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 abstract int getDependencyOptimizationValue ()
Control dependency optimizations (default is ON). Remove redundant control dependencies, which may enable other optimization.
.tensorflow.RewriterConfig.Toggle dependency_optimization = 8;
public abstract boolean getDisableMetaOptimizer ()
Disable the entire meta optimizer (off by default).
bool disable_meta_optimizer = 19;
public abstract boolean getDisableModelPruning ()
If true, don't remove unnecessary ops from the graph
bool disable_model_pruning = 2;
public abstract 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 abstract 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;
public abstract RewriterConfig.Toggle getFunctionOptimization ()
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
public abstract int getFunctionOptimizationValue ()
Function optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle function_optimization = 10;
public abstract RewriterConfig.Toggle getImplementationSelector ()
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
public abstract int getImplementationSelectorValue ()
Enable the swap of kernel implementations based on the device placement (default is ON).
.tensorflow.RewriterConfig.Toggle implementation_selector = 22;
public abstract VerifierConfig getInterOptimizerVerifierConfig ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public abstract VerifierConfigOrBuilder getInterOptimizerVerifierConfigOrBuilder ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public abstract 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 abstract 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;
public abstract RewriterConfig.Toggle getLoopOptimization ()
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
public abstract int getLoopOptimizationValue ()
Loop optimizations (default is ON).
.tensorflow.RewriterConfig.Toggle loop_optimization = 9;
abstrato 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 abstract 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 abstract 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 abstract 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;
abstratos públicas 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 abstract int getMetaOptimizerIterationsValue ()
Controls how many times we run the optimizers in meta optimizer (default is once).
.tensorflow.RewriterConfig.NumIterationsType meta_optimizer_iterations = 12;
public abstract 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 abstract 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 abstract 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 abstract 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 abstract 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 abstract List <String> 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;
public abstract RewriterConfig.Toggle getPinToHostOptimization ()
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
public abstract int getPinToHostOptimizationValue ()
Force small ops onto the CPU (default is OFF).
.tensorflow.RewriterConfig.Toggle pin_to_host_optimization = 18;
public abstract 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 abstract VerifierConfigOrBuilder getPostOptimizationVerifierConfigOrBuilder ()
VerifierConfig specifying the verifiers to be run at the end, after all optimizers have run.
.tensorflow.VerifierConfig post_optimization_verifier_config = 301;
public abstract RewriterConfig.Toggle getRemapping ()
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
public abstract int getRemappingValue ()
Remapping (default is ON) Remap subgraphs onto more efficient implementations.
.tensorflow.RewriterConfig.Toggle remapping = 14;
public abstract 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 abstract 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 abstract ScopedAllocatorOptions getScopedAllocatorOpts ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public abstract ScopedAllocatorOptionsOrBuilder getScopedAllocatorOptsOrBuilder ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;
public abstract RewriterConfig.Toggle getShapeOptimization ()
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;
public abstract int getShapeOptimizationValue ()
Shape optimizations (default is ON) Simplify computations made on shapes.
.tensorflow.RewriterConfig.Toggle shape_optimization = 13;
public abstract 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 abstract boolean hasInterOptimizerVerifierConfig ()
VerifierConfig specifying the verifiers to be run after every optimizer.
.tensorflow.VerifierConfig inter_optimizer_verifier_config = 300;
public abstract 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 abstract boolean hasScopedAllocatorOpts ()
.tensorflow.ScopedAllocatorOptions scoped_allocator_opts = 16;