lớp cuối cùng tĩnh công khai ConfigProto.Builder
Session configuration parameters. The system picks appropriate values for fields that are not set.Loại
tensorflow.ConfigProto của Protobuf.ConfigProto Phương pháp công khai
| Cấu hìnhProto.Builder | addAllDeviceFilters (giá trị Iterable<String>) When any filters are present sessions will ignore all devices which do not match the filters. |
| Cấu hìnhProto.Builder | addAllSessionInterOpThreadPool (Iterable<? mở rộng các giá trị ThreadPoolOptionProto >) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | addDeviceFilters (Giá trị chuỗi) When any filters are present sessions will ignore all devices which do not match the filters. |
| Cấu hìnhProto.Builder | addDeviceFiltersBytes (giá trị com.google.protobuf.ByteString) When any filters are present sessions will ignore all devices which do not match the filters. |
| Cấu hìnhProto.Builder | addRepeatedField (trường com.google.protobuf.Descriptors.FieldDescriptor, Giá trị đối tượng) |
| Cấu hìnhProto.Builder | addSessionInterOpThreadPool (giá trị ThreadPoolOptionProto ) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | addSessionInterOpThreadPool (chỉ mục int, giá trị ThreadPoolOptionProto ) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | addSessionInterOpThreadPool ( ThreadPoolOptionProto.Builder builderForValue) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | addSessionInterOpThreadPool (chỉ mục int, ThreadPoolOptionProto.Builder builderForValue) This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProto.Builder | addSessionInterOpThreadPoolBuilder (chỉ mục 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. |
| Cấu hìnhProto | xây dựng () |
| Cấu hìnhProto | |
| Cấu hìnhProto.Builder | thông thoáng () |
| Cấu hìnhProto.Builder | clearAllowSoftPlacement () Whether soft placement is allowed. |
| Cấu hìnhProto.Builder | clearClusterDef () Optional list of all workers to use in this session. |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | ClearDeviceFilters () When any filters are present sessions will ignore all devices which do not match the filters. |
| Cấu hìnhProto.Builder | rõ ràngThử nghiệm () .tensorflow.ConfigProto.Experimental experimental = 16; |
| Cấu hìnhProto.Builder | ClearField (trường com.google.protobuf.Descriptors.FieldDescriptor) |
| Cấu hìnhProto.Builder | ClearGpuOptions () Options that apply to all GPUs. |
| Cấu hìnhProto.Builder | ClearGraphOptions () Options that apply to all graphs. |
| Cấu hìnhProto.Builder | clearInterOpParallelismThreads () Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. |
| Cấu hìnhProto.Builder | clearIntraOpParallelismThreads () The execution of an individual op (for some op types) can be parallelized on a pool of intra_op_parallelism_threads. |
| Cấu hìnhProto.Builder | clearIsolateSessionState () If true, any resources such as Variables used in the session will not be shared with other sessions. |
| Cấu hìnhProto.Builder | clearLogDevicePlacement () Whether device placements should be logged. |
| Cấu hìnhProto.Builder | clearOneof (com.google.protobuf.Descriptors.OneofDescriptor oneof) |
| Cấu hìnhProto.Builder | clearOperationTimeoutInMs () Global timeout for all blocking operations in this session. |
| Cấu hìnhProto.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). |
| Cấu hìnhProto.Builder | ClearRpcOptions () Options that apply when this session uses the distributed runtime. |
| Cấu hìnhProto.Builder | clearSessionInterOpThreadPool () This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | clearShareClusterDevicesInSession () When true, WorkerSessions are created with device attributes from the full cluster. |
| Cấu hìnhProto.Builder | clearUsePerSessionThreads () If true, use a new set of threads for this session rather than the global pool of threads. |
| Cấu hìnhProto.Builder | dòng vô tính () |
| boolean | chứaDeviceCount (Khóa chuỗi) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| boolean | getAllowSoftPlacement () Whether soft placement is allowed. |
| ClusterDef | getClusterDef () Optional list of all workers to use in this session. |
| ClusterDef.Builder | getClusterDefBuilder () Optional list of all workers to use in this session. |
| ClusterDefOrBuilder | getClusterDefOrBuilder () Optional list of all workers to use in this session. |
| Cấu hìnhProto | |
| com.google.protobuf.Descriptors.Descriptor tĩnh cuối cùng | |
| com.google.protobuf.Descriptors.Descriptor | |
| Bản đồ<Chuỗi, Số nguyên> | getDeviceCount () Thay vào đó hãy sử dụng getDeviceCountMap() . |
| int | getDeviceCountCount () Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Bản đồ<Chuỗi, Số nguyên> | getDeviceCountMap () Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| int | getDeviceCountOrDefault (Khóa chuỗi, int defaultValue) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| int | getDeviceCountOrThrow (Khóa chuỗi) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Sợi dây | getDeviceFilters (chỉ mục int) When any filters are present sessions will ignore all devices which do not match the filters. |
| com.google.protobuf.ByteString | getDeviceFiltersBytes (chỉ mục int) When any filters are present sessions will ignore all devices which do not match the filters. |
| int | 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.Experimental | lấyThử nghiệm () .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.Experimental.Builder | getExperimentalBuilder () .tensorflow.ConfigProto.Experimental experimental = 16; |
| ConfigProto.ExperimentalOrBuilder | getExperimentalOrBuilder () .tensorflow.ConfigProto.Experimental experimental = 16; |
| Tùy chọn GPU | getGpuOptions () Options that apply to all GPUs. |
| GPUOptions.Builder | getGpuOptionsBuilder () Options that apply to all GPUs. |
| GPUOptionsOrBuilder | getGpuOptionsOrBuilder () Options that apply to all GPUs. |
| Tùy chọn đồ thị | getGraphOptions () Options that apply to all graphs. |
| GraphOptions.Builder | getGraphOptionsBuilder () Options that apply to all graphs. |
| GraphOptionsOrBuilder | getGraphOptionsOrBuilder () Options that apply to all graphs. |
| int | getInterOpParallelismThreads () Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. |
| int | getIntraOpParallelismThreads () The execution of an individual op (for some op types) can be parallelized on a pool of intra_op_parallelism_threads. |
| boolean | getIsolateSessionState () If true, any resources such as Variables used in the session will not be shared with other sessions. |
| boolean | getLogDevicePlacement () Whether device placements should be logged. |
| Bản đồ<Chuỗi, Số nguyên> | getMutableDeviceCount () Thay vào đó hãy sử dụng các bộ truy cập đột biến thay thế. |
| dài | getOperationTimeoutInMs () Global timeout for all blocking operations in this session. |
| 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). |
| Tùy chọn RPC | 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 (chỉ mục int) This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProto.Builder | getSessionInterOpThreadPoolBuilder (chỉ mục int) This option is experimental - it may be replaced with a different mechanism in the future. |
| Danh sách< ThreadPoolOptionProto.Builder > | getSessionInterOpThreadPoolBuilderList () This option is experimental - it may be replaced with a different mechanism in the future. |
| int | getSessionInterOpThreadPoolCount () This option is experimental - it may be replaced with a different mechanism in the future. |
| Danh sách< ThreadPoolOptionProto > | getSessionInterOpThreadPoolList () This option is experimental - it may be replaced with a different mechanism in the future. |
| ThreadPoolOptionProtoOrBuilder | getSessionInterOpThreadPoolOrBuilder (chỉ mục int) This option is experimental - it may be replaced with a different mechanism in the future. |
| Danh sách<? mở rộng ThreadPoolOptionProtoOrBuilder > | getSessionInterOpThreadPoolOrBuilderList () This option is experimental - it may be replaced with a different mechanism in the future. |
| boolean | getShareClusterDevicesInSession () When true, WorkerSessions are created with device attributes from the full cluster. |
| boolean | getUsePerSessionThreads () If true, use a new set of threads for this session rather than the global pool of threads. |
| boolean | hasClusterDef () Optional list of all workers to use in this session. |
| boolean | cóThử nghiệm () .tensorflow.ConfigProto.Experimental experimental = 16; |
| boolean | hasGpuOptions () Options that apply to all GPUs. |
| boolean | hasGraphOptions () Options that apply to all graphs. |
| boolean | hasRpcOptions () Options that apply when this session uses the distributed runtime. |
| boolean cuối cùng | |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | hợp nhấtExperimental ( ConfigProto.Experimental value) .tensorflow.ConfigProto.Experimental experimental = 16; |
| Cấu hìnhProto.Builder | mergeFrom (com.google.protobuf.Message other) |
| Cấu hìnhProto.Builder | mergeFrom (đầu vào com.google.protobuf.CodedInputStream, com.google.protobuf.ExtensionRegistryLite ExtensionRegistry) |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | mergeRpcOptions (giá trị RPCOptions ) Options that apply when this session uses the distributed runtime. |
| ConfigProto.Builder cuối cùng | hợp nhấtUnknownFields (com.google.protobuf.UnknownFieldSet knownFields) |
| Cấu hìnhProto.Builder | putAllDeviceCount (Giá trị bản đồ<String, Integer>) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Cấu hìnhProto.Builder | putDeviceCount (Khóa chuỗi, giá trị int) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Cấu hìnhProto.Builder | RemoveDeviceCount (Khóa chuỗi) Map from device type name (e.g., "CPU" or "GPU" ) to maximum number of devices of that type to use. |
| Cấu hìnhProto.Builder | RemoveSessionInterOpThreadPool (chỉ mục int) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | setAllowSoftPlacement (giá trị boolean) Whether soft placement is allowed. |
| Cấu hìnhProto.Builder | setClusterDef ( ClusterDef.Builder builderForValue) Optional list of all workers to use in this session. |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | setDeviceFilters (chỉ mục int, giá trị chuỗi) When any filters are present sessions will ignore all devices which do not match the filters. |
| Cấu hìnhProto.Builder | setExperimental ( ConfigProto.Experimental value) .tensorflow.ConfigProto.Experimental experimental = 16; |
| Cấu hìnhProto.Builder | setExperimental ( ConfigProto.Experimental.Builder builderForValue) .tensorflow.ConfigProto.Experimental experimental = 16; |
| Cấu hìnhProto.Builder | setField (trường com.google.protobuf.Descriptors.FieldDescriptor, Giá trị đối tượng) |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | |
| Cấu hìnhProto.Builder | setInterOpParallelismThreads (giá trị int) Nodes that perform blocking operations are enqueued on a pool of inter_op_parallelism_threads available in each process. |
| Cấu hìnhProto.Builder | setIntraOpParallelismThreads (giá trị int) The execution of an individual op (for some op types) can be parallelized on a pool of intra_op_parallelism_threads. |
| Cấu hìnhProto.Builder | setIsolateSessionState (giá trị boolean) If true, any resources such as Variables used in the session will not be shared with other sessions. |
| Cấu hìnhProto.Builder | setLogDevicePlacement (giá trị boolean) Whether device placements should be logged. |
| Cấu hìnhProto.Builder | setOperationTimeoutInMs (giá trị dài) Global timeout for all blocking operations in this session. |
| Cấu hìnhProto.Builder | setPlacementPeriod (giá trị int) 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). |
| Cấu hìnhProto.Builder | setRepeatedField (trường com.google.protobuf.Descriptors.FieldDescriptor, chỉ mục int, giá trị đối tượng) |
| Cấu hìnhProto.Builder | setRpcOptions (giá trị RPCOptions ) Options that apply when this session uses the distributed runtime. |
| Cấu hìnhProto.Builder | setRpcOptions ( RPCOptions.Builder builderForValue) Options that apply when this session uses the distributed runtime. |
| Cấu hìnhProto.Builder | setSessionInterOpThreadPool (chỉ mục int, giá trị ThreadPoolOptionProto ) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | setSessionInterOpThreadPool (chỉ mục int, ThreadPoolOptionProto.Builder builderForValue) This option is experimental - it may be replaced with a different mechanism in the future. |
| Cấu hìnhProto.Builder | setShareClusterDevicesInSession (giá trị boolean) When true, WorkerSessions are created with device attributes from the full cluster. |
| ConfigProto.Builder cuối cùng | setUnknownFields (com.google.protobuf.UnknownFieldSet knownFields) |
| Cấu hìnhProto.Builder | setUsePerSessionThreads (giá trị boolean) If true, use a new set of threads for this session rather than the global pool of threads. |
Phương pháp kế thừa
Phương pháp công khai
public ConfigProto.Builder addAllDeviceFilters (giá trị 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; công khai ConfigProto.Builder addAllSessionInterOpThreadPool (các giá trị Iterable<? mở rộng 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; công khai ConfigProto.Builder addDeviceFilters (Giá trị chuỗi)
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; ConfigProto.Builder công khai addDeviceFiltersBytes (giá trị 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; công khai ConfigProto.Builder addRepeatedField (trường com.google.protobuf.Descriptors.FieldDescriptor, Giá trị đối tượng)
công khai ConfigProto.Builder addSessionInterOpThreadPool (giá trị 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 (chỉ mục int, giá trị 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; công khai 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; ConfigProto.Builder công khai addSessionInterOpThreadPool (chỉ mục 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 (chỉ mục 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; công khai 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; công khai 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; công khai ConfigProto.Builder clearClusterDef ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; công khai 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; công khai ConfigProto.Builder clearExperimental ()
.tensorflow.ConfigProto.Experimental experimental = 16;
công khai ConfigProto.Builder clearGpuOptions ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; công khai 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; công khai 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; công khai ConfigProto.Builder clearLogDevicePlacement ()
Whether device placements should be logged.
bool log_device_placement = 8; public 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; công khai 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; công khai ConfigProto.Builder clearRpcOptions ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; công khai 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; công khai 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; công khai 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; boolean công khai chứaDeviceCount (Khóa chuỗi)
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; boolean công khai 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; công khai ClusterDef getClusterDef ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; công khai ClusterDef.Builder getClusterDefBuilder ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; công khai ClusterDefOrBuilder getClusterDefOrBuilder ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; công khai tĩnh cuối cùng com.google.protobuf.Descriptors.Descriptor getDescriptor ()
com.google.protobuf.Descriptors.Descriptor công khai getDescriptorForType ()
int công khai 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; Bản đồ công khai<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 (Khóa chuỗi, 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 (Khóa chuỗi)
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; Chuỗi công khai getDeviceFilters (chỉ mục 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; com.google.protobuf.ByteString getDeviceFiltersBytes công khai (chỉ mục 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; int công khai 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; com.google.protobuf.ProtocolStringList công khai 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; công khai ConfigProto.Experimental getExperimental ()
.tensorflow.ConfigProto.Experimental experimental = 16;
ConfigProto.Experimental.Builder công khai getExperimentalBuilder ()
.tensorflow.ConfigProto.Experimental experimental = 16;
ConfigProto.ExperimentalOrBuilder công khai getExperimentalOrBuilder ()
.tensorflow.ConfigProto.Experimental experimental = 16;
GPUOptions công khai getGpuOptions ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; GPUOptions.Builder công khai getGpuOptionsBuilder ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; GPUOptionsOrBuilder công khai getGpuOptionsOrBuilder ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; GraphOptions công khai getGraphOptions ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; GraphOptions.Builder công khai getGraphOptionsBuilder ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; GraphOptionsOrBuilder công khai 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; boolean công khai 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; boolean công khai getLogDevicePlacement ()
Whether device placements should be logged.
bool log_device_placement = 8; Bản đồ công khai<String, Integer> getMutableDeviceCount ()
Thay vào đó hãy sử dụng các bộ truy cập đột biến thay thế.
công khai long 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; int công khai 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; RPCOptions công khai getRpcOptions ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; RPCOptions.Builder công khai getRpcOptionsBuilder ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; RPCOptionsOrBuilder công khai getRpcOptionsOrBuilder ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; public ThreadPoolOptionProto getSessionInterOpThreadPool (chỉ mục 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;công khai ThreadPoolOptionProto.Builder getSessionInterOpThreadPoolBuilder (chỉ mục 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; Danh sách công khai< 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; int công khai 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; Danh sách công khai< 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; công khai ThreadPoolOptionProtoOrBuilder getSessionInterOpThreadPoolOrBuilder (chỉ mục 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; Danh sách công khai<? mở rộng 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; boolean công khai 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; boolean công khai 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; boolean công khai hasClusterDef ()
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; boolean công khai hasExperimental ()
.tensorflow.ConfigProto.Experimental experimental = 16;
boolean công khai hasGpuOptions ()
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; boolean công khai hasGraphOptions ()
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; boolean công khai hasRpcOptions ()
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; boolean cuối cùng công khai được khởi tạo ()
công khai ConfigProto.Builder mergeClusterDef (giá trị ClusterDef )
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; công khai ConfigProto.Builder mergeExperimental (giá trị ConfigProto.Experimental )
.tensorflow.ConfigProto.Experimental experimental = 16;
công khai ConfigProto.Builder mergeFrom (đầu vào com.google.protobuf.CodedInputStream, com.google.protobuf.ExtensionRegistryLite ExtensionRegistry)
Ném
| IOException |
|---|
công khai ConfigProto.Builder mergeGpuOptions (giá trị GPUOptions )
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; công khai ConfigProto.Builder mergeGraphOptions (giá trị GraphOptions )
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; công khai ConfigProto.Builder mergeRpcOptions (giá trị RPCOptions )
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; bản cuối cùng công khai ConfigProto.Builder mergeUnknownFields (com.google.protobuf.UnknownFieldSet knownFields)
công khai ConfigProto.Builder putAllDeviceCount (giá trị Bản đồ<Chuỗi, Số nguyên>)
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 (Khóa chuỗi, giá trị 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; công khai ConfigProto.Builder RemoveDeviceCount (Khóa chuỗi)
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; ConfigProto.Builder công khai xóaSessionInterOpThreadPool (chỉ mục 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 (giá trị boolean)
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; ConfigProto.Builder setClusterDef công khai ( ClusterDef.Builder builderForValue)
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14; ConfigProto.Builder setClusterDef công khai (giá trị ClusterDef )
Optional list of all workers to use in this session.
.tensorflow.ClusterDef cluster_def = 14;public ConfigProto.Builder setDeviceFilters (chỉ mục int, giá trị chuỗi)
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; công khai ConfigProto.Builder setExperimental (giá trị ConfigProto.Experimental )
.tensorflow.ConfigProto.Experimental experimental = 16;
công khai ConfigProto.Builder setExperimental ( ConfigProto.Experimental.Builder builderForValue)
.tensorflow.ConfigProto.Experimental experimental = 16;
ConfigProto.Builder setField công khai (trường com.google.protobuf.Descriptors.FieldDescriptor, Giá trị đối tượng)
ConfigProto.Builder setGpuOptions công khai ( GPUOptions.Builder builderForValue)
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; ConfigProto.Builder setGpuOptions công khai (giá trị GPUOptions )
Options that apply to all GPUs.
.tensorflow.GPUOptions gpu_options = 6; ConfigProto.Builder setGraphOptions công khai ( GraphOptions.Builder builderForValue)
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public ConfigProto.Builder setGraphOptions (giá trị GraphOptions )
Options that apply to all graphs.
.tensorflow.GraphOptions graph_options = 10; public ConfigProto.Builder setInterOpParallelismThreads (giá trị int)
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 (giá trị int)
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; ConfigProto.Builder setIsolateSessionState công khai (giá trị boolean)
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; ConfigProto.Builder setLogDevicePlacement công khai (giá trị boolean)
Whether device placements should be logged.
bool log_device_placement = 8; ConfigProto.Builder setOperationTimeoutInMs công khai (giá trị dài)
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; ConfigProto.Builder setPlacementPeriod công khai (giá trị int)
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; ConfigProto.Builder setRepeatedField công khai (trường com.google.protobuf.Descriptors.FieldDescriptor, chỉ mục int, giá trị đối tượng)
ConfigProto.Builder setRpcOptions công khai (giá trị RPCOptions )
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; ConfigProto.Builder setRpcOptions công khai ( RPCOptions.Builder builderForValue)
Options that apply when this session uses the distributed runtime.
.tensorflow.RPCOptions rpc_options = 13; public ConfigProto.Builder setSessionInterOpThreadPool (chỉ mục int, giá trị 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; ConfigProto.Builder setSessionInterOpThreadPool công khai (chỉ mục 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; ConfigProto.Builder setShareClusterDevicesInSession công khai (giá trị boolean)
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; trận chung kết công khai ConfigProto.Builder setUnknownFields (com.google.protobuf.UnknownFieldSet knownFields)
ConfigProto.Builder setUsePerSessionThreads công khai (giá trị boolean)
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;