Everything inside Experimental is subject to change and is not subject to API stability guarantees in https://www.tensorflow.org/guide/version_compat.
tensorflow.RunOptions.Experimental
Public Methods
RunOptions.Experimental.Builder |
addRepeatedField(com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
|
RunOptions.Experimental |
build()
|
RunOptions.Experimental | |
RunOptions.Experimental.Builder |
clear()
|
RunOptions.Experimental.Builder |
clearCollectiveGraphKey()
If non-zero, declares that this graph is going to use collective ops and must synchronize step_ids with any other graph with this same group_key value (in a distributed computation where tasks run disjoint graphs). |
RunOptions.Experimental.Builder |
clearField(com.google.protobuf.Descriptors.FieldDescriptor field)
|
RunOptions.Experimental.Builder |
clearOneof(com.google.protobuf.Descriptors.OneofDescriptor oneof)
|
RunOptions.Experimental.Builder |
clearRunHandlerPoolOptions()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
RunOptions.Experimental.Builder |
clearUseRunHandlerPool()
If true, then operations (using the inter-op pool) across all session::run() calls will be centrally scheduled, optimizing for (median and tail) latency. |
RunOptions.Experimental.Builder |
clone()
|
long |
getCollectiveGraphKey()
If non-zero, declares that this graph is going to use collective ops and must synchronize step_ids with any other graph with this same group_key value (in a distributed computation where tasks run disjoint graphs). |
RunOptions.Experimental | |
final static com.google.protobuf.Descriptors.Descriptor | |
com.google.protobuf.Descriptors.Descriptor | |
RunOptions.Experimental.RunHandlerPoolOptions |
getRunHandlerPoolOptions()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
RunOptions.Experimental.RunHandlerPoolOptions.Builder |
getRunHandlerPoolOptionsBuilder()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
RunOptions.Experimental.RunHandlerPoolOptionsOrBuilder |
getRunHandlerPoolOptionsOrBuilder()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
boolean |
getUseRunHandlerPool()
If true, then operations (using the inter-op pool) across all session::run() calls will be centrally scheduled, optimizing for (median and tail) latency. |
boolean |
hasRunHandlerPoolOptions()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
final boolean | |
RunOptions.Experimental.Builder |
mergeFrom(com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
|
RunOptions.Experimental.Builder |
mergeFrom(com.google.protobuf.Message other)
|
RunOptions.Experimental.Builder |
mergeRunHandlerPoolOptions(RunOptions.Experimental.RunHandlerPoolOptions value)
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
final RunOptions.Experimental.Builder |
mergeUnknownFields(com.google.protobuf.UnknownFieldSet unknownFields)
|
RunOptions.Experimental.Builder |
setCollectiveGraphKey(long value)
If non-zero, declares that this graph is going to use collective ops and must synchronize step_ids with any other graph with this same group_key value (in a distributed computation where tasks run disjoint graphs). |
RunOptions.Experimental.Builder |
setField(com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
|
RunOptions.Experimental.Builder |
setRepeatedField(com.google.protobuf.Descriptors.FieldDescriptor field, int index, Object value)
|
RunOptions.Experimental.Builder |
setRunHandlerPoolOptions(RunOptions.Experimental.RunHandlerPoolOptions value)
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
RunOptions.Experimental.Builder |
setRunHandlerPoolOptions(RunOptions.Experimental.RunHandlerPoolOptions.Builder builderForValue)
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
|
final RunOptions.Experimental.Builder |
setUnknownFields(com.google.protobuf.UnknownFieldSet unknownFields)
|
RunOptions.Experimental.Builder |
setUseRunHandlerPool(boolean value)
If true, then operations (using the inter-op pool) across all session::run() calls will be centrally scheduled, optimizing for (median and tail) latency. |
Inherited Methods
Public Methods
public RunOptions.Experimental.Builder addRepeatedField (com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
public RunOptions.Experimental.Builder clearCollectiveGraphKey ()
If non-zero, declares that this graph is going to use collective ops and must synchronize step_ids with any other graph with this same group_key value (in a distributed computation where tasks run disjoint graphs).
int64 collective_graph_key = 1;
public RunOptions.Experimental.Builder clearField (com.google.protobuf.Descriptors.FieldDescriptor field)
public RunOptions.Experimental.Builder clearOneof (com.google.protobuf.Descriptors.OneofDescriptor oneof)
public RunOptions.Experimental.Builder clearRunHandlerPoolOptions ()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public RunOptions.Experimental.Builder clearUseRunHandlerPool ()
If true, then operations (using the inter-op pool) across all session::run() calls will be centrally scheduled, optimizing for (median and tail) latency. Consider using this option for CPU-bound workloads like inference.
bool use_run_handler_pool = 2;
public long getCollectiveGraphKey ()
If non-zero, declares that this graph is going to use collective ops and must synchronize step_ids with any other graph with this same group_key value (in a distributed computation where tasks run disjoint graphs).
int64 collective_graph_key = 1;
public static final com.google.protobuf.Descriptors.Descriptor getDescriptor ()
public com.google.protobuf.Descriptors.Descriptor getDescriptorForType ()
public RunOptions.Experimental.RunHandlerPoolOptions getRunHandlerPoolOptions ()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public RunOptions.Experimental.RunHandlerPoolOptions.Builder getRunHandlerPoolOptionsBuilder ()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public RunOptions.Experimental.RunHandlerPoolOptionsOrBuilder getRunHandlerPoolOptionsOrBuilder ()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public boolean getUseRunHandlerPool ()
If true, then operations (using the inter-op pool) across all session::run() calls will be centrally scheduled, optimizing for (median and tail) latency. Consider using this option for CPU-bound workloads like inference.
bool use_run_handler_pool = 2;
public boolean hasRunHandlerPoolOptions ()
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public final boolean isInitialized ()
public RunOptions.Experimental.Builder mergeFrom (com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
Throws
IOException |
---|
public RunOptions.Experimental.Builder mergeRunHandlerPoolOptions (RunOptions.Experimental.RunHandlerPoolOptions value)
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public final RunOptions.Experimental.Builder mergeUnknownFields (com.google.protobuf.UnknownFieldSet unknownFields)
public RunOptions.Experimental.Builder setCollectiveGraphKey (long value)
If non-zero, declares that this graph is going to use collective ops and must synchronize step_ids with any other graph with this same group_key value (in a distributed computation where tasks run disjoint graphs).
int64 collective_graph_key = 1;
public RunOptions.Experimental.Builder setField (com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
public RunOptions.Experimental.Builder setRepeatedField (com.google.protobuf.Descriptors.FieldDescriptor field, int index, Object value)
public RunOptions.Experimental.Builder setRunHandlerPoolOptions (RunOptions.Experimental.RunHandlerPoolOptions value)
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public RunOptions.Experimental.Builder setRunHandlerPoolOptions (RunOptions.Experimental.RunHandlerPoolOptions.Builder builderForValue)
.tensorflow.RunOptions.Experimental.RunHandlerPoolOptions run_handler_pool_options = 3;
public final RunOptions.Experimental.Builder setUnknownFields (com.google.protobuf.UnknownFieldSet unknownFields)
public RunOptions.Experimental.Builder setUseRunHandlerPool (boolean value)
If true, then operations (using the inter-op pool) across all session::run() calls will be centrally scheduled, optimizing for (median and tail) latency. Consider using this option for CPU-bound workloads like inference.
bool use_run_handler_pool = 2;