public static final class
CallableOptions.Builder
Defines a subgraph in another `GraphDef` as a set of feed points and nodes to be fetched or executed. Compare with the arguments to `Session::Run()`.
tensorflow.CallableOptions
Public Methods
CallableOptions.Builder |
addAllFeed(Iterable<String> values)
Tensors to be fed in the callable. |
CallableOptions.Builder |
addAllFetch(Iterable<String> values)
Fetches. |
CallableOptions.Builder |
addAllTarget(Iterable<String> values)
Target Nodes. |
CallableOptions.Builder |
addAllTensorConnection(Iterable<? extends TensorConnection> values)
Tensors to be connected in the callable. |
CallableOptions.Builder |
addFeed(String value)
Tensors to be fed in the callable. |
CallableOptions.Builder |
addFeedBytes(com.google.protobuf.ByteString value)
Tensors to be fed in the callable. |
CallableOptions.Builder |
addFetch(String value)
Fetches. |
CallableOptions.Builder |
addFetchBytes(com.google.protobuf.ByteString value)
Fetches. |
CallableOptions.Builder |
addRepeatedField(com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
|
CallableOptions.Builder |
addTarget(String value)
Target Nodes. |
CallableOptions.Builder |
addTargetBytes(com.google.protobuf.ByteString value)
Target Nodes. |
CallableOptions.Builder | |
CallableOptions.Builder |
addTensorConnection(int index, TensorConnection.Builder builderForValue)
Tensors to be connected in the callable. |
CallableOptions.Builder | |
CallableOptions.Builder |
addTensorConnection(TensorConnection.Builder builderForValue)
Tensors to be connected in the callable. |
TensorConnection.Builder |
addTensorConnectionBuilder()
Tensors to be connected in the callable. |
TensorConnection.Builder |
addTensorConnectionBuilder(int index)
Tensors to be connected in the callable. |
CallableOptions |
build()
|
CallableOptions | |
CallableOptions.Builder |
clear()
|
CallableOptions.Builder |
clearFeed()
Tensors to be fed in the callable. |
CallableOptions.Builder | |
CallableOptions.Builder |
clearFetch()
Fetches. |
CallableOptions.Builder | |
CallableOptions.Builder |
clearFetchSkipSync()
By default, RunCallable() will synchronize the GPU stream before returning fetched tensors on a GPU device, to ensure that the values in those tensors have been produced. |
CallableOptions.Builder |
clearField(com.google.protobuf.Descriptors.FieldDescriptor field)
|
CallableOptions.Builder |
clearOneof(com.google.protobuf.Descriptors.OneofDescriptor oneof)
|
CallableOptions.Builder |
clearRunOptions()
Options that will be applied to each run. |
CallableOptions.Builder |
clearTarget()
Target Nodes. |
CallableOptions.Builder |
clearTensorConnection()
Tensors to be connected in the callable. |
CallableOptions.Builder |
clone()
|
boolean |
containsFeedDevices(String key)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
boolean |
containsFetchDevices(String key)
map<string, string> fetch_devices = 7;
|
CallableOptions | |
final static com.google.protobuf.Descriptors.Descriptor | |
com.google.protobuf.Descriptors.Descriptor | |
String |
getFeed(int index)
Tensors to be fed in the callable. |
com.google.protobuf.ByteString |
getFeedBytes(int index)
Tensors to be fed in the callable. |
int |
getFeedCount()
Tensors to be fed in the callable. |
Map<String, String> |
getFeedDevices()
Use
getFeedDevicesMap() instead. |
int |
getFeedDevicesCount()
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
Map<String, String> |
getFeedDevicesMap()
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
String |
getFeedDevicesOrDefault(String key, String defaultValue)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
String |
getFeedDevicesOrThrow(String key)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
com.google.protobuf.ProtocolStringList |
getFeedList()
Tensors to be fed in the callable. |
String |
getFetch(int index)
Fetches. |
com.google.protobuf.ByteString |
getFetchBytes(int index)
Fetches. |
int |
getFetchCount()
Fetches. |
Map<String, String> |
getFetchDevices()
Use
getFetchDevicesMap() instead. |
int |
getFetchDevicesCount()
map<string, string> fetch_devices = 7;
|
Map<String, String> |
getFetchDevicesMap()
map<string, string> fetch_devices = 7;
|
String |
getFetchDevicesOrDefault(String key, String defaultValue)
map<string, string> fetch_devices = 7;
|
String |
getFetchDevicesOrThrow(String key)
map<string, string> fetch_devices = 7;
|
com.google.protobuf.ProtocolStringList |
getFetchList()
Fetches. |
boolean |
getFetchSkipSync()
By default, RunCallable() will synchronize the GPU stream before returning fetched tensors on a GPU device, to ensure that the values in those tensors have been produced. |
Map<String, String> |
getMutableFeedDevices()
Use alternate mutation accessors instead.
|
Map<String, String> |
getMutableFetchDevices()
Use alternate mutation accessors instead.
|
RunOptions |
getRunOptions()
Options that will be applied to each run. |
RunOptions.Builder |
getRunOptionsBuilder()
Options that will be applied to each run. |
RunOptionsOrBuilder |
getRunOptionsOrBuilder()
Options that will be applied to each run. |
String |
getTarget(int index)
Target Nodes. |
com.google.protobuf.ByteString |
getTargetBytes(int index)
Target Nodes. |
int |
getTargetCount()
Target Nodes. |
com.google.protobuf.ProtocolStringList |
getTargetList()
Target Nodes. |
TensorConnection |
getTensorConnection(int index)
Tensors to be connected in the callable. |
TensorConnection.Builder |
getTensorConnectionBuilder(int index)
Tensors to be connected in the callable. |
List<TensorConnection.Builder> |
getTensorConnectionBuilderList()
Tensors to be connected in the callable. |
int |
getTensorConnectionCount()
Tensors to be connected in the callable. |
List<TensorConnection> |
getTensorConnectionList()
Tensors to be connected in the callable. |
TensorConnectionOrBuilder |
getTensorConnectionOrBuilder(int index)
Tensors to be connected in the callable. |
List<? extends TensorConnectionOrBuilder> |
getTensorConnectionOrBuilderList()
Tensors to be connected in the callable. |
boolean |
hasRunOptions()
Options that will be applied to each run. |
final boolean | |
CallableOptions.Builder |
mergeFrom(com.google.protobuf.Message other)
|
CallableOptions.Builder |
mergeFrom(com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
|
CallableOptions.Builder | |
final CallableOptions.Builder |
mergeUnknownFields(com.google.protobuf.UnknownFieldSet unknownFields)
|
CallableOptions.Builder |
putAllFeedDevices(Map<String, String> values)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
CallableOptions.Builder |
putAllFetchDevices(Map<String, String> values)
map<string, string> fetch_devices = 7;
|
CallableOptions.Builder |
putFeedDevices(String key, String value)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
CallableOptions.Builder |
putFetchDevices(String key, String value)
map<string, string> fetch_devices = 7;
|
CallableOptions.Builder |
removeFeedDevices(String key)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. |
CallableOptions.Builder |
removeFetchDevices(String key)
map<string, string> fetch_devices = 7;
|
CallableOptions.Builder |
removeTensorConnection(int index)
Tensors to be connected in the callable. |
CallableOptions.Builder |
setFeed(int index, String value)
Tensors to be fed in the callable. |
CallableOptions.Builder |
setFetch(int index, String value)
Fetches. |
CallableOptions.Builder |
setFetchSkipSync(boolean value)
By default, RunCallable() will synchronize the GPU stream before returning fetched tensors on a GPU device, to ensure that the values in those tensors have been produced. |
CallableOptions.Builder |
setField(com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
|
CallableOptions.Builder |
setRepeatedField(com.google.protobuf.Descriptors.FieldDescriptor field, int index, Object value)
|
CallableOptions.Builder | |
CallableOptions.Builder | |
CallableOptions.Builder |
setTarget(int index, String value)
Target Nodes. |
CallableOptions.Builder | |
CallableOptions.Builder |
setTensorConnection(int index, TensorConnection.Builder builderForValue)
Tensors to be connected in the callable. |
final CallableOptions.Builder |
setUnknownFields(com.google.protobuf.UnknownFieldSet unknownFields)
|
Inherited Methods
Public Methods
public CallableOptions.Builder addAllFeed (Iterable<String> values)
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public CallableOptions.Builder addAllFetch (Iterable<String> values)
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public CallableOptions.Builder addAllTarget (Iterable<String> values)
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public CallableOptions.Builder addAllTensorConnection (Iterable<? extends TensorConnection> values)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder addFeed (String value)
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public CallableOptions.Builder addFeedBytes (com.google.protobuf.ByteString value)
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public CallableOptions.Builder addFetch (String value)
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public CallableOptions.Builder addFetchBytes (com.google.protobuf.ByteString value)
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public CallableOptions.Builder addRepeatedField (com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
public CallableOptions.Builder addTarget (String value)
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public CallableOptions.Builder addTargetBytes (com.google.protobuf.ByteString value)
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public CallableOptions.Builder addTensorConnection (int index, TensorConnection value)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder addTensorConnection (int index, TensorConnection.Builder builderForValue)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder addTensorConnection (TensorConnection value)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder addTensorConnection (TensorConnection.Builder builderForValue)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public TensorConnection.Builder addTensorConnectionBuilder ()
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public TensorConnection.Builder addTensorConnectionBuilder (int index)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder clearFeed ()
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public CallableOptions.Builder clearFetch ()
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public CallableOptions.Builder clearFetchSkipSync ()
By default, RunCallable() will synchronize the GPU stream before returning fetched tensors on a GPU device, to ensure that the values in those tensors have been produced. This simplifies interacting with the tensors, but potentially incurs a performance hit. If this options is set to true, the caller is responsible for ensuring that the values in the fetched tensors have been produced before they are used. The caller can do this by invoking `Device::Sync()` on the underlying device(s), or by feeding the tensors back to the same Session using `feed_devices` with the same corresponding device name.
bool fetch_skip_sync = 8;
public CallableOptions.Builder clearRunOptions ()
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public CallableOptions.Builder clearTarget ()
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public CallableOptions.Builder clearTensorConnection ()
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public boolean containsFeedDevices (String key)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public boolean containsFetchDevices (String key)
map<string, string> fetch_devices = 7;
public static final com.google.protobuf.Descriptors.Descriptor getDescriptor ()
public com.google.protobuf.Descriptors.Descriptor getDescriptorForType ()
public String getFeed (int index)
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public com.google.protobuf.ByteString getFeedBytes (int index)
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public int getFeedCount ()
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public int getFeedDevicesCount ()
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public Map<String, String> getFeedDevicesMap ()
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public String getFeedDevicesOrDefault (String key, String defaultValue)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public String getFeedDevicesOrThrow (String key)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public com.google.protobuf.ProtocolStringList getFeedList ()
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public String getFetch (int index)
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public com.google.protobuf.ByteString getFetchBytes (int index)
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public int getFetchCount ()
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public int getFetchDevicesCount ()
map<string, string> fetch_devices = 7;
public Map<String, String> getFetchDevicesMap ()
map<string, string> fetch_devices = 7;
public String getFetchDevicesOrDefault (String key, String defaultValue)
map<string, string> fetch_devices = 7;
public String getFetchDevicesOrThrow (String key)
map<string, string> fetch_devices = 7;
public com.google.protobuf.ProtocolStringList getFetchList ()
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public boolean getFetchSkipSync ()
By default, RunCallable() will synchronize the GPU stream before returning fetched tensors on a GPU device, to ensure that the values in those tensors have been produced. This simplifies interacting with the tensors, but potentially incurs a performance hit. If this options is set to true, the caller is responsible for ensuring that the values in the fetched tensors have been produced before they are used. The caller can do this by invoking `Device::Sync()` on the underlying device(s), or by feeding the tensors back to the same Session using `feed_devices` with the same corresponding device name.
bool fetch_skip_sync = 8;
public Map<String, String> getMutableFeedDevices ()
Use alternate mutation accessors instead.
public Map<String, String> getMutableFetchDevices ()
Use alternate mutation accessors instead.
public RunOptions getRunOptions ()
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public RunOptions.Builder getRunOptionsBuilder ()
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public RunOptionsOrBuilder getRunOptionsOrBuilder ()
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public String getTarget (int index)
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public com.google.protobuf.ByteString getTargetBytes (int index)
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public int getTargetCount ()
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public com.google.protobuf.ProtocolStringList getTargetList ()
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public TensorConnection getTensorConnection (int index)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public TensorConnection.Builder getTensorConnectionBuilder (int index)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public List<TensorConnection.Builder> getTensorConnectionBuilderList ()
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public int getTensorConnectionCount ()
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public List<TensorConnection> getTensorConnectionList ()
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public TensorConnectionOrBuilder getTensorConnectionOrBuilder (int index)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public List<? extends TensorConnectionOrBuilder> getTensorConnectionOrBuilderList ()
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public boolean hasRunOptions ()
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public final boolean isInitialized ()
public CallableOptions.Builder mergeFrom (com.google.protobuf.CodedInputStream input, com.google.protobuf.ExtensionRegistryLite extensionRegistry)
Throws
IOException |
---|
public CallableOptions.Builder mergeRunOptions (RunOptions value)
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public final CallableOptions.Builder mergeUnknownFields (com.google.protobuf.UnknownFieldSet unknownFields)
public CallableOptions.Builder putAllFeedDevices (Map<String, String> values)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public CallableOptions.Builder putAllFetchDevices (Map<String, String> values)
map<string, string> fetch_devices = 7;
public CallableOptions.Builder putFeedDevices (String key, String value)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public CallableOptions.Builder putFetchDevices (String key, String value)
map<string, string> fetch_devices = 7;
public CallableOptions.Builder removeFeedDevices (String key)
The Tensor objects fed in the callable and fetched from the callable are expected to be backed by host (CPU) memory by default. The options below allow changing that - feeding tensors backed by device memory, or returning tensors that are backed by device memory. The maps below map the name of a feed/fetch tensor (which appears in 'feed' or 'fetch' fields above), to the fully qualified name of the device owning the memory backing the contents of the tensor. For example, creating a callable with the following options: CallableOptions { feed: "a:0" feed: "b:0" fetch: "x:0" fetch: "y:0" feed_devices: { "a:0": "/job:localhost/replica:0/task:0/device:GPU:0" } fetch_devices: { "y:0": "/job:localhost/replica:0/task:0/device:GPU:0" } } means that the Callable expects: - The first argument ("a:0") is a Tensor backed by GPU memory. - The second argument ("b:0") is a Tensor backed by host memory. and of its return values: - The first output ("x:0") will be backed by host memory. - The second output ("y:0") will be backed by GPU memory. FEEDS: It is the responsibility of the caller to ensure that the memory of the fed tensors will be correctly initialized and synchronized before it is accessed by operations executed during the call to Session::RunCallable(). This is typically ensured by using the TensorFlow memory allocators (Device::GetAllocator()) to create the Tensor to be fed. Alternatively, for CUDA-enabled GPU devices, this typically means that the operation that produced the contents of the tensor has completed, i.e., the CUDA stream has been synchronized (e.g., via cuCtxSynchronize() or cuStreamSynchronize()).
map<string, string> feed_devices = 6;
public CallableOptions.Builder removeFetchDevices (String key)
map<string, string> fetch_devices = 7;
public CallableOptions.Builder removeTensorConnection (int index)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder setFeed (int index, String value)
Tensors to be fed in the callable. Each feed is the name of a tensor.
repeated string feed = 1;
public CallableOptions.Builder setFetch (int index, String value)
Fetches. A list of tensor names. The caller of the callable expects a tensor to be returned for each fetch[i] (see RunStepResponse.tensor). The order of specified fetches does not change the execution order.
repeated string fetch = 2;
public CallableOptions.Builder setFetchSkipSync (boolean value)
By default, RunCallable() will synchronize the GPU stream before returning fetched tensors on a GPU device, to ensure that the values in those tensors have been produced. This simplifies interacting with the tensors, but potentially incurs a performance hit. If this options is set to true, the caller is responsible for ensuring that the values in the fetched tensors have been produced before they are used. The caller can do this by invoking `Device::Sync()` on the underlying device(s), or by feeding the tensors back to the same Session using `feed_devices` with the same corresponding device name.
bool fetch_skip_sync = 8;
public CallableOptions.Builder setField (com.google.protobuf.Descriptors.FieldDescriptor field, Object value)
public CallableOptions.Builder setRepeatedField (com.google.protobuf.Descriptors.FieldDescriptor field, int index, Object value)
public CallableOptions.Builder setRunOptions (RunOptions.Builder builderForValue)
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public CallableOptions.Builder setRunOptions (RunOptions value)
Options that will be applied to each run.
.tensorflow.RunOptions run_options = 4;
public CallableOptions.Builder setTarget (int index, String value)
Target Nodes. A list of node names. The named nodes will be run by the callable but their outputs will not be returned.
repeated string target = 3;
public CallableOptions.Builder setTensorConnection (int index, TensorConnection value)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;
public CallableOptions.Builder setTensorConnection (int index, TensorConnection.Builder builderForValue)
Tensors to be connected in the callable. Each TensorConnection denotes a pair of tensors in the graph, between which an edge will be created in the callable.
repeated .tensorflow.TensorConnection tensor_connection = 5;