общедоступный статический конечный класс ConfigProto.Builder
Session configuration parameters. The system picks appropriate values for fields that are not set.
tensorflow.ConfigProto Публичные методы
| ConfigProto.Builder | addAllDeviceFilters (значения Iterable<String>) When any filters are present sessions will ignore all devices which do not match the filters. |
| ConfigProto.Builder | addAllSessionInterOpThreadPool (Iterable<? расширяет значения ThreadPoolOptionProto >) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | addDeviceFilters (строковое значение) When any filters are present sessions will ignore all devices which do not match the filters. |
| ConfigProto.Builder | addDeviceFiltersBytes (значение com.google.protobuf.ByteString) When any filters are present sessions will ignore all devices which do not match the filters. |
| ConfigProto.Builder | addRepeatedField (поле com.google.protobuf.Descriptors.FieldDescriptor, значение объекта) |
| ConfigProto.Builder | addSessionInterOpThreadPool (значение ThreadPoolOptionProto ) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | addSessionInterOpThreadPool (индекс int, значение ThreadPoolOptionProto ) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | addSessionInterOpThreadPool ( ThreadPoolOptionProto.Builder builderForValue) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | addSessionInterOpThreadPool (индекс int, ThreadPoolOptionProto.Builder builderForValue) This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProto.Builder | addSessionInterOpThreadPoolBuilder (индекс int) This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProto.Builder | addSessionInterOpThreadPoolBuilder () This option is experimental - it may be replaced with a different mechanism in the future. |
| КонфигПрото | строить () |
| КонфигПрото | |
| ConfigProto.Builder | прозрачный () |
| ConfigProto.Builder | ClearAllowSoftPlacement () Whether soft placement is allowed. |
| ConfigProto.Builder | очиститьКластерДеф () Optional list of all workers to use in this session. |
| ConfigProto.Builder | |
| ConfigProto.Builder | ОчиститьДевицеФильтерс () When any filters are present sessions will ignore all devices which do not match the filters. |
| ConfigProto.Builder | очиститьЭкспериментальный () .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.Builder | ClearField (поле com.google.protobuf.Descriptors.FieldDescriptor) |
| ConfigProto.Builder | ОчиститьГпуОпционс () Options that apply to all GPUs. |
| ConfigProto.Builder | ClearGraphOptions () Options that apply to all graphs. |
| ConfigProto.Builder | ОчиститьИнтерОпПараллелизмТредс () Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. |
| ConfigProto.Builder | очиститьIntraOpParallelismThreads () The execution of an individual op (for some op types) can be parallelized on a pool of intra_op_parallelism_threads. |
| ConfigProto.Builder | ClearIsolateSessionState () If true, any resources such as Variables used in the session will not be shared with other sessions. |
| ConfigProto.Builder | яснологдевицеплацемент () Whether device placements should be logged. |
| ConfigProto.Builder | ClearOneof (com.google.protobuf.Descriptors.OneofDescriptor oneof) |
| ConfigProto.Builder | ClearOperationTimeoutInMs () Global timeout for all blocking operations in this session. |
| ConfigProto.Builder | ОчиститьПлацементПериод () Assignment of Nodes to Devices is recomputed every placement_period steps until the system warms up (at which point the recomputation typically slows down automatically). |
| ConfigProto.Builder | очиститьRpcOptions () Options that apply when this session uses the distributed runtime. |
| ConfigProto.Builder | ClearSessionInterOpThreadPool () This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | ClearShareClusterDevicesInSession () When true, WorkerSessions are created with device attributes from the full cluster. |
| ConfigProto.Builder | очиститьUsePerSessionThreads () If true, use a new set of threads for this session rather than the global pool of threads. |
| ConfigProto.Builder | клон () |
| логическое значение | содержитDeviceCount (строковый ключ) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| логическое значение | getAllowSoftPlacement () Whether soft placement is allowed. |
| КластерДеф | getClusterDef () Optional list of all workers to use in this session. |
| ClusterDef.Builder | getClusterDefBuilder () Optional list of all workers to use in this session. |
| КластерДефОрБилдер | getClusterDefOrBuilder () Optional list of all workers to use in this session. |
| КонфигПрото | |
| окончательный статический com.google.protobuf.Descriptors.Descriptor | |
| com.google.protobuf.Descriptors.Descriptor | |
| Карта<String, Integer> | getDeviceCount () Вместо этого используйте getDeviceCountMap() . |
| интервал | getDeviceCountCount () Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Карта<String, Integer> | getDeviceCountMap () Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| интервал | getDeviceCountOrDefault (строковый ключ, int defaultValue) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| интервал | getDeviceCountOrThrow (строковый ключ) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Нить | getDeviceFilters (индекс целого числа) When any filters are present sessions will ignore all devices which do not match the filters. |
| com.google.protobuf.ByteString | getDeviceFiltersBytes (индекс целого числа) When any filters are present sessions will ignore all devices which do not match the filters. |
| интервал | getDeviceFiltersCount () When any filters are present sessions will ignore all devices which do not match the filters. |
| com.google.protobuf.ProtocolStringList | getDeviceFiltersList () When any filters are present sessions will ignore all devices which do not match the filters. |
| ConfigProto.Экспериментальный | получитьЭкспериментальный () .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.Experimental.Builder | getExperimentalBuilder () .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.ExperimentalOrBuilder | getExperimentalOrBuilder () .tensorflow.ConfigProto.Experimental experimental = 16; |
| Параметры графического процессора | получитьGpuOptions () Options that apply to all GPUs. |
| GPUOptions.Builder | getGpuOptionsBuilder () Options that apply to all GPUs. |
| GPUOptionsOrBuilder | getGpuOptionsOrBuilder () Options that apply to all GPUs. |
| Параметры графика | getGraphOptions () Options that apply to all graphs. |
| GraphOptions.Builder | getGraphOptionsBuilder () Options that apply to all graphs. |
| GraphOptionsOrBuilder | getGraphOptionsOrBuilder () Options that apply to all graphs. |
| интервал | getInterOpParallelismThreads () Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. |
| интервал | getIntraOpParallelismThreads () The execution of an individual op (for some op types) can be parallelized on a pool of intra_op_parallelism_threads. |
| логическое значение | getIsolateSessionState () If true, any resources such as Variables used in the session will not be shared with other sessions. |
| логическое значение | getLogDevicePlacement () Whether device placements should be logged. |
| Карта<String, Integer> | getMutableDeviceCount () Вместо этого используйте альтернативные методы доступа к мутациям. |
| длинный | getOperationTimeoutInMs () Global timeout for all blocking operations in this session. |
| интервал | getPlacementPeriod () Assignment of Nodes to Devices is recomputed every placement_period steps until the system warms up (at which point the recomputation typically slows down automatically). |
| RPCOptions | getRpcOptions () Options that apply when this session uses the distributed runtime. |
| RPCOptions.Builder | getRpcOptionsBuilder () Options that apply when this session uses the distributed runtime. |
| RPCOptionsOrBuilder | getRpcOptionsOrBuilder () Options that apply when this session uses the distributed runtime. |
| ThreadPoolOptionProto | getSessionInterOpThreadPool (индекс int) This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProto.Builder | getSessionInterOpThreadPoolBuilder (индекс int) This option is experimental - it may be replaced with a different mechanism in the future. |
| Список < ThreadPoolOptionProto.Builder > | getSessionInterOpThreadPoolBuilderList () This option is experimental - it may be replaced with a different mechanism in the future. |
| интервал | getSessionInterOpThreadPoolCount () This option is experimental - it may be replaced with a different mechanism in the future. |
| Список < ThreadPoolOptionProto > | getSessionInterOpThreadPoolList () This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProtoOrBuilder | getSessionInterOpThreadPoolOrBuilder (индекс int) This option is experimental - it may be replaced with a different mechanism in the future. |
| Список <? расширяет ThreadPoolOptionProtoOrBuilder > | getSessionInterOpThreadPoolOrBuilderList () This option is experimental - it may be replaced with a different mechanism in the future. |
| логическое значение | getShareClusterDevicesInSession () When true, WorkerSessions are created with device attributes from the full cluster. |
| логическое значение | getUsePerSessionThreads () If true, use a new set of threads for this session rather than the global pool of threads. |
| логическое значение | имеетКластерДеф () Optional list of all workers to use in this session. |
| логическое значение | имеетЭкспериментальный () .tensorflow.ConfigProto.Experimental experimental = 16; |
| логическое значение | имеетGpuOptions () Options that apply to all GPUs. |
| логическое значение | имеетграфопционы () Options that apply to all graphs. |
| логическое значение | имеетRpcOptions () Options that apply when this session uses the distributed runtime. |
| последнее логическое значение | |
| ConfigProto.Builder | |
| ConfigProto.Builder | mergeExperimental (значение ConfigProto.Experimental ) .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.Builder | mergeFrom (com.google.protobuf.Message другое) |
| ConfigProto.Builder | mergeFrom (ввод com.google.protobuf.CodedInputStream, com.google.protobuf.ExtensionRegistryLite ExtensionRegistry) |
| ConfigProto.Builder | |
| ConfigProto.Builder | |
| ConfigProto.Builder | mergeRpcOptions (значение RPCOptions ) Options that apply when this session uses the distributed runtime. |
| окончательный ConfigProto.Builder | mergeUnknownFields (com.google.protobuf.UnknownFieldSetknownFields) |
| ConfigProto.Builder | putAllDeviceCount (значения Map<String, Integer>) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| ConfigProto.Builder | putDeviceCount (строковый ключ, целое значение) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| ConfigProto.Builder | RemoveDeviceCount (строковый ключ) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| ConfigProto.Builder | RemoveSessionInterOpThreadPool (индекс int) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | setAllowSoftPlacement (логическое значение) Whether soft placement is allowed. |
| ConfigProto.Builder | setClusterDef ( ClusterDef.Builder builderForValue) Optional list of all workers to use in this session. |
| ConfigProto.Builder | |
| ConfigProto.Builder | setDeviceFilters (индекс int, строковое значение) When any filters are present sessions will ignore all devices which do not match the filters. |
| ConfigProto.Builder | setExperimental (значение ConfigProto.Experimental ) .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.Builder | setExperimental ( ConfigProto.Experimental.Builder builderForValue) .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.Builder | setField (поле com.google.protobuf.Descriptors.FieldDescriptor, значение объекта) |
| ConfigProto.Builder | |
| ConfigProto.Builder | |
| ConfigProto.Builder | |
| ConfigProto.Builder | |
| ConfigProto.Builder | setInterOpParallelismThreads (целое значение) Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. |
| ConfigProto.Builder | setIntraOpParallelismThreads (целое значение) The execution of an individual op (for some op types) can be parallelized on a pool of intra_op_parallelism_threads. |
| ConfigProto.Builder | setIsolateSessionState (логическое значение) If true, any resources such as Variables used in the session will not be shared with other sessions. |
| ConfigProto.Builder | setLogDevicePlacement (логическое значение) Whether device placements should be logged. |
| ConfigProto.Builder | setOperationTimeoutInMs (длинное значение) Global timeout for all blocking operations in this session. |
| ConfigProto.Builder | setPlacementPeriod (целое значение) Assignment of Nodes to Devices is recomputed every placement_period steps until the system warms up (at which point the recomputation typically slows down automatically). |
| ConfigProto.Builder | setRepeatedField (поле com.google.protobuf.Descriptors.FieldDescriptor, индекс int, значение объекта) |
| ConfigProto.Builder | setRpcOptions (значение RPCOptions ) Options that apply when this session uses the distributed runtime. |
| ConfigProto.Builder | setRpcOptions ( RPCOptions.Builder builderForValue) Options that apply when this session uses the distributed runtime. |
| ConfigProto.Builder | setSessionInterOpThreadPool (индекс int, значение ThreadPoolOptionProto ) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | setSessionInterOpThreadPool (индекс int, ThreadPoolOptionProto.Builder builderForValue) This option is experimental - it may be replaced with a different mechanism in the future. |
| ConfigProto.Builder | setShareClusterDevicesInSession (логическое значение) When true, WorkerSessions are created with device attributes from the full cluster. |
| окончательный ConfigProto.Builder | setUnknownFields (com.google.protobuf.UnknownFieldSetknownFields) |
| ConfigProto.Builder | setUsePerSessionThreads (логическое значение) If true, use a new set of threads for this session rather than the global pool of threads. |
Унаследованные методы
Публичные методы
public ConfigProto.Builder addAllDeviceFilters (значения Iterable<String>)
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public ConfigProto.Builder addAllSessionInterOpThreadPool (Iterable<? расширяет значения ThreadPoolOptionProto >)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder addDeviceFilters (строковое значение)
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public ConfigProto.Builder addDeviceFiltersBytes (значение com.google.protobuf.ByteString)
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public ConfigProto.Builder addRepeatedField (поле com.google.protobuf.Descriptors.FieldDescriptor, значение объекта)
public ConfigProto.Builder addSessionInterOpThreadPool (значение ThreadPoolOptionProto )
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder addSessionInterOpThreadPool (индекс int, значение ThreadPoolOptionProto )
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder addSessionInterOpThreadPool ( ThreadPoolOptionProto.Builder builderForValue)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder addSessionInterOpThreadPool (индекс int, ThreadPoolOptionProto.Builder builderForValue)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ThreadPoolOptionProto.Builder addSessionInterOpThreadPoolBuilder (индекс int)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступный ThreadPoolOptionProto.Builder addSessionInterOpThreadPoolBuilder ()
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступный ConfigProto.Builder ClearAllowSoftPlacement ()
Whether soft placement is allowed. If allow_soft_placement is true, an op will be placed on CPU if 1. there's no GPU implementation for the OP or 2. no GPU devices are known or registered or 3. need to co-locate with reftype input(s) which are from CPU.
bool allow_soft_placement = 7; общедоступный ConfigProto.Builder ClearClusterDef ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; общедоступный ConfigProto.Builder ClearDeviceFilters ()
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; общедоступный ConfigProto.Builder ClearExperimental ()
.tensorflow.ConfigProto.Experimental experimental = 16;
общедоступный ConfigProto.Builder ClearGpuOptions ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; общедоступный ConfigProto.Builder ClearGraphOptions ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public ConfigProto.Builder ClearInterOpParallelismThreads ()
Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. 0 means the system picks an appropriate number. Negative means all operations are performed in caller's thread. Note that the first Session created in the process sets the number of threads for all future sessions unless use_per_session_threads is true or session_inter_op_thread_pool is configured.
int32 inter_op_parallelism_threads = 5; public ConfigProto.Builder ClearIntraOpParallelismThreads ()
The execution of an individual op (for some op types) can be
parallelized on a pool of intra_op_parallelism_threads.
0 means the system picks an appropriate number.
If you create an ordinary session, e.g., from Python or C++,
then there is exactly one intra op thread pool per process.
The first session created determines the number of threads in this pool.
All subsequent sessions reuse/share this one global pool.
There are notable exceptions to the default behavior describe above:
1. There is an environment variable for overriding this thread pool,
named TF_OVERRIDE_GLOBAL_THREADPOOL.
2. When connecting to a server, such as a remote `tf.train.Server`
instance, then this option will be ignored altogether.
int32 intra_op_parallelism_threads = 2; public ConfigProto.Builder ClearIsolateSessionState ()
If true, any resources such as Variables used in the session will not be shared with other sessions. However, when clusterspec propagation is enabled, this field is ignored and sessions are always isolated.
bool isolate_session_state = 15; общедоступный ConfigProto.Builder ClearLogDevicePlacement ()
Whether device placements should be logged.
bool log_device_placement = 8; общедоступный ConfigProto.Builder ClearOneof (com.google.protobuf.Descriptors.OneofDescriptor oneof)
общедоступный ConfigProto.Builder ClearOperationTimeoutInMs ()
Global timeout for all blocking operations in this session. If non-zero, and not overridden on a per-operation basis, this value will be used as the deadline for all blocking operations.
int64 operation_timeout_in_ms = 11; общедоступный ConfigProto.Builder ClearPlacementPeriod ()
Assignment of Nodes to Devices is recomputed every placement_period steps until the system warms up (at which point the recomputation typically slows down automatically).
int32 placement_period = 3; общедоступный ConfigProto.Builder ClearRpcOptions ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; общедоступный ConfigProto.Builder ClearSessionInterOpThreadPool ()
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступный ConfigProto.Builder ClearShareClusterDevicesInSession ()
When true, WorkerSessions are created with device attributes from the full cluster. This is helpful when a worker wants to partition a graph (for example during a PartitionedCallOp).
bool share_cluster_devices_in_session = 17; общедоступный ConfigProto.Builder ClearUsePerSessionThreads ()
If true, use a new set of threads for this session rather than the global pool of threads. Only supported by direct sessions. If false, use the global threads created by the first session, or the per-session thread pools configured by session_inter_op_thread_pool. This option is deprecated. The same effect can be achieved by setting session_inter_op_thread_pool to have one element, whose num_threads equals inter_op_parallelism_threads.
bool use_per_session_threads = 9; общедоступное логическое значение containsDeviceCount (строковый ключ)
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; общедоступное логическое значение getAllowSoftPlacement ()
Whether soft placement is allowed. If allow_soft_placement is true, an op will be placed on CPU if 1. there's no GPU implementation for the OP or 2. no GPU devices are known or registered or 3. need to co-locate with reftype input(s) which are from CPU.
bool allow_soft_placement = 7; общественный ClusterDef getClusterDef ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; общедоступный ClusterDef.Builder getClusterDefBuilder ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; public ClusterDefOrBuilder getClusterDefOrBuilder ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; общедоступный статический окончательный com.google.protobuf.Descriptors.Descriptor getDescriptor ()
общедоступный com.google.protobuf.Descriptors.Descriptor getDescriptorForType ()
общедоступная карта<String, Integer> getDeviceCount ()
Вместо этого используйте getDeviceCountMap() .
public int getDeviceCountCount ()
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; общедоступная карта<String, Integer> getDeviceCountMap ()
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; public int getDeviceCountOrDefault (строковый ключ, int defaultValue)
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; public int getDeviceCountOrThrow (строковый ключ)
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; общедоступная строка getDeviceFilters (индекс int)
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public com.google.protobuf.ByteString getDeviceFiltersBytes (индекс int)
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public int getDeviceFiltersCount ()
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public com.google.protobuf.ProtocolStringList getDeviceFiltersList ()
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; общедоступный ConfigProto.Experimental getExperimental ()
.tensorflow.ConfigProto.Experimental experimental = 16;
общедоступный ConfigProto.Experimental.Builder getExperimentalBuilder ()
.tensorflow.ConfigProto.Experimental experimental = 16;
общедоступный ConfigProto.ExperimentalOrBuilder getExperimentalOrBuilder ()
.tensorflow.ConfigProto.Experimental experimental = 16;
общедоступные GPUOptions getGpuOptions ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; общедоступный GPUOptions.Builder getGpuOptionsBuilder ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; общедоступный GPUOptionsOrBuilder getGpuOptionsOrBuilder ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; общедоступные GraphOptions getGraphOptions ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; общедоступный GraphOptions.Builder getGraphOptionsBuilder ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; общедоступный GraphOptionsOrBuilder getGraphOptionsOrBuilder ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public int getInterOpParallelismThreads ()
Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. 0 means the system picks an appropriate number. Negative means all operations are performed in caller's thread. Note that the first Session created in the process sets the number of threads for all future sessions unless use_per_session_threads is true or session_inter_op_thread_pool is configured.
int32 inter_op_parallelism_threads = 5; public int getIntraOpParallelismThreads ()
The execution of an individual op (for some op types) can be
parallelized on a pool of intra_op_parallelism_threads.
0 means the system picks an appropriate number.
If you create an ordinary session, e.g., from Python or C++,
then there is exactly one intra op thread pool per process.
The first session created determines the number of threads in this pool.
All subsequent sessions reuse/share this one global pool.
There are notable exceptions to the default behavior describe above:
1. There is an environment variable for overriding this thread pool,
named TF_OVERRIDE_GLOBAL_THREADPOOL.
2. When connecting to a server, such as a remote `tf.train.Server`
instance, then this option will be ignored altogether.
int32 intra_op_parallelism_threads = 2; общедоступное логическое значение getIsolateSessionState ()
If true, any resources such as Variables used in the session will not be shared with other sessions. However, when clusterspec propagation is enabled, this field is ignored and sessions are always isolated.
bool isolate_session_state = 15; общедоступное логическое значение getLogDevicePlacement ()
Whether device placements should be logged.
bool log_device_placement = 8; общедоступная карта <String, Integer> getMutableDeviceCount ()
Вместо этого используйте альтернативные методы доступа к мутациям.
общедоступный длинный getOperationTimeoutInMs ()
Global timeout for all blocking operations in this session. If non-zero, and not overridden on a per-operation basis, this value will be used as the deadline for all blocking operations.
int64 operation_timeout_in_ms = 11; public int getPlacementPeriod ()
Assignment of Nodes to Devices is recomputed every placement_period steps until the system warms up (at which point the recomputation typically slows down automatically).
int32 placement_period = 3; общедоступные RPCOptions getRpcOptions ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; общедоступный RPCOptions.Builder getRpcOptionsBuilder ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; общедоступный RPCOptionsOrBuilder getRpcOptionsOrBuilder ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; public ThreadPoolOptionProto getSessionInterOpThreadPool (индекс int)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ThreadPoolOptionProto.Builder getSessionInterOpThreadPoolBuilder (индекс int)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступный список < ThreadPoolOptionProto.Builder > getSessionInterOpThreadPoolBuilderList ()
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public int getSessionInterOpThreadPoolCount ()
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступный список < ThreadPoolOptionProto > getSessionInterOpThreadPoolList ()
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ThreadPoolOptionProtoOrBuilder getSessionInterOpThreadPoolOrBuilder (индекс int)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступный список<? расширяет ThreadPoolOptionProtoOrBuilder > getSessionInterOpThreadPoolOrBuilderList ()
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; общедоступное логическое значение getShareClusterDevicesInSession ()
When true, WorkerSessions are created with device attributes from the full cluster. This is helpful when a worker wants to partition a graph (for example during a PartitionedCallOp).
bool share_cluster_devices_in_session = 17; общедоступное логическое значение getUsePerSessionThreads ()
If true, use a new set of threads for this session rather than the global pool of threads. Only supported by direct sessions. If false, use the global threads created by the first session, or the per-session thread pools configured by session_inter_op_thread_pool. This option is deprecated. The same effect can be achieved by setting session_inter_op_thread_pool to have one element, whose num_threads equals inter_op_parallelism_threads.
bool use_per_session_threads = 9; общедоступное логическое значение hasClusterDef ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; общедоступное логическое значение hasExperimental ()
.tensorflow.ConfigProto.Experimental experimental = 16;
общедоступное логическое значение hasGpuOptions ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; общедоступное логическое значение hasGraphOptions ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; общедоступное логическое значение hasRpcOptions ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; публичное финальное логическое значение isInitialized ()
public ConfigProto.Builder mergeClusterDef (значение ClusterDef )
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; public ConfigProto.Builder mergeExperimental (значение ConfigProto.Experimental )
.tensorflow.ConfigProto.Experimental experimental = 16;
public ConfigProto.Builder mergeFrom (вход com.google.protobuf.CodedInputStream, com.google.protobuf.ExtensionRegistryLite ExtensionRegistry)
Броски
| Исключение IO |
|---|
public ConfigProto.Builder mergeGpuOptions (значение GPUOptions )
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; public ConfigProto.Builder mergeGraphOptions (значение GraphOptions )
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public ConfigProto.Builder mergeRpcOptions (значение RPCOptions )
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; общедоступный окончательный ConfigProto.Builder mergeUnknownFields (com.google.protobuf.UnknownFieldSetknownFields)
public ConfigProto.Builder putAllDeviceCount (значения Map<String, Integer>)
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; public ConfigProto.Builder putDeviceCount (строковый ключ, значение int)
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; общедоступный ConfigProto.Builder RemoveDeviceCount (строковый ключ)
Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. If a particular device type is not found in the map, the system picks an appropriate number.
map<string, int32> device_count = 1; public ConfigProto.Builder RemoveSessionInterOpThreadPool (индекс int)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder setAllowSoftPlacement (логическое значение)
Whether soft placement is allowed. If allow_soft_placement is true, an op will be placed on CPU if 1. there's no GPU implementation for the OP or 2. no GPU devices are known or registered or 3. need to co-locate with reftype input(s) which are from CPU.
bool allow_soft_placement = 7; public ConfigProto.Builder setClusterDef ( ClusterDef.Builder builderForValue)
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; public ConfigProto.Builder setClusterDef (значение ClusterDef )
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; public ConfigProto.Builder setDeviceFilters (индекс int, строковое значение)
When any filters are present sessions will ignore all devices which do not match the filters. Each filter can be partially specified, e.g. "/job:ps" "/job:worker/replica:3", etc.
repeated string device_filters = 4; public ConfigProto.Builder setExperimental (значение ConfigProto.Experimental )
.tensorflow.ConfigProto.Experimental experimental = 16;
общедоступный ConfigProto.Builder setExperimental ( ConfigProto.Experimental.Builder builderForValue)
.tensorflow.ConfigProto.Experimental experimental = 16;
public ConfigProto.Builder setField (поле com.google.protobuf.Descriptors.FieldDescriptor, значение объекта)
public ConfigProto.Builder setGpuOptions ( GPUOptions.Builder builderForValue)
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; public ConfigProto.Builder setGpuOptions (значение GPUOptions )
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; общедоступный ConfigProto.Builder setGraphOptions ( GraphOptions.Builder builderForValue)
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public ConfigProto.Builder setGraphOptions (значение GraphOptions )
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public ConfigProto.Builder setInterOpParallelismThreads (целое значение)
Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. 0 means the system picks an appropriate number. Negative means all operations are performed in caller's thread. Note that the first Session created in the process sets the number of threads for all future sessions unless use_per_session_threads is true or session_inter_op_thread_pool is configured.
int32 inter_op_parallelism_threads = 5; public ConfigProto.Builder setIntraOpParallelismThreads (целое значение)
The execution of an individual op (for some op types) can be
parallelized on a pool of intra_op_parallelism_threads.
0 means the system picks an appropriate number.
If you create an ordinary session, e.g., from Python or C++,
then there is exactly one intra op thread pool per process.
The first session created determines the number of threads in this pool.
All subsequent sessions reuse/share this one global pool.
There are notable exceptions to the default behavior describe above:
1. There is an environment variable for overriding this thread pool,
named TF_OVERRIDE_GLOBAL_THREADPOOL.
2. When connecting to a server, such as a remote `tf.train.Server`
instance, then this option will be ignored altogether.
int32 intra_op_parallelism_threads = 2; public ConfigProto.Builder setIsolateSessionState (логическое значение)
If true, any resources such as Variables used in the session will not be shared with other sessions. However, when clusterspec propagation is enabled, this field is ignored and sessions are always isolated.
bool isolate_session_state = 15; public ConfigProto.Builder setLogDevicePlacement (логическое значение)
Whether device placements should be logged.
bool log_device_placement = 8; public ConfigProto.Builder setOperationTimeoutInMs (длинное значение)
Global timeout for all blocking operations in this session. If non-zero, and not overridden on a per-operation basis, this value will be used as the deadline for all blocking operations.
int64 operation_timeout_in_ms = 11; public ConfigProto.Builder setPlacementPeriod (целое значение)
Assignment of Nodes to Devices is recomputed every placement_period steps until the system warms up (at which point the recomputation typically slows down automatically).
int32 placement_period = 3; public ConfigProto.Builder setRepeatedField (поле com.google.protobuf.Descriptors.FieldDescriptor, индекс int, значение объекта)
public ConfigProto.Builder setRpcOptions (значение RPCOptions )
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; общедоступный ConfigProto.Builder setRpcOptions ( RPCOptions.Builder builderForValue)
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; public ConfigProto.Builder setSessionInterOpThreadPool (индекс int, значение ThreadPoolOptionProto )
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder setSessionInterOpThreadPool (индекс int, ThreadPoolOptionProto.Builder builderForValue)
This option is experimental - it may be replaced with a different mechanism in the future. Configures session thread pools. If this is configured, then RunOptions for a Run call can select the thread pool to use. The intended use is for when some session invocations need to run in a background pool limited to a small number of threads: - For example, a session may be configured to have one large pool (for regular compute) and one small pool (for periodic, low priority work); using the small pool is currently the mechanism for limiting the inter-op parallelism of the low priority work. Note that it does not limit the parallelism of work spawned by a single op kernel implementation. - Using this setting is normally not needed in training, but may help some serving use cases. - It is also generally recommended to set the global_name field of this proto, to avoid creating multiple large pools. It is typically better to run the non-low-priority work, even across sessions, in a single large pool.
repeated .tensorflow.ThreadPoolOptionProto session_inter_op_thread_pool = 12; public ConfigProto.Builder setShareClusterDevicesInSession (логическое значение)
When true, WorkerSessions are created with device attributes from the full cluster. This is helpful when a worker wants to partition a graph (for example during a PartitionedCallOp).
bool share_cluster_devices_in_session = 17; общедоступный окончательный ConfigProto.Builder setUnknownFields (com.google.protobuf.UnknownFieldSetknownFields)
public ConfigProto.Builder setUsePerSessionThreads (логическое значение)
If true, use a new set of threads for this session rather than the global pool of threads. Only supported by direct sessions. If false, use the global threads created by the first session, or the per-session thread pools configured by session_inter_op_thread_pool. This option is deprecated. The same effect can be achieved by setting session_inter_op_thread_pool to have one element, whose num_threads equals inter_op_parallelism_threads.
bool use_per_session_threads = 9;